Characteristics of patients diagnosed with pancreatic cancer who access palliative care: An observational study

Palliative care can help patients with pancreatic cancer manage their symptoms and improve their quality of life. Although palliative care should be available to all patients with pancreatic cancer, we don’t know whether patients are getting this support early enough. This study aimed to understand whether certain groups of patients with pancreatic cancer are more or less likely to present to palliative care with high symptom distress or problems. By understanding which groups of patients present to palliative care with high symptom distress or problems, we can design strategies to help these patients get support earlier. We found that the majority of patients with pancreatic cancer first saw palliative care when their overall health had worsened. Only a small number of patients who saw a palliative care service included in this study lived in a regional or remote area. Patients who presented to community palliative care generally had high symptom distress and problems. Compared to English-speaking patients, a greater proportion of patients who did not speak English were deceased at the end of their first presentation and their families/carers were more likely to present with high distress. These findings indicate that patients with pancreatic cancer may not be accessing palliative care early enough. It also inspires more research on ways to make palliative care easier to access in the community and for patients who live in regional and remote areas and do not speak English.

The majority of patients with pancreatic cancer are diagnosed when the cancer has metastasised [1]. Coupled with a poor prognosis, patients with pancreatic cancer experience a high symptom burden [2‐4], which, if inadequately managed, contributes to a poor quality of life (QoL) [5] as well as poorer survival outcomes [6]. The World Health Organisation defines palliative care as “an approach that improves the quality of life of patients and their families faced with a life-threatening illness, through the prevention and relief of suffering by means of early identification and impeccable assessment and treatment of pain and other problems, physical, psychosocial and spiritual” [7]. Clinical management pathways for pancreatic cancer recommend that all patients be offered a referral for PC assessment [8] and that early PC occur concurrently with active cancer treatment [9].

Engagement with PC services has been demonstrated to have numerous benefits. A systematic review and meta-analysis of 43 randomised controlled trials exploring the impact of PC on patient outcomes found PC to be associated with statistically and clinically significant improvements in QoL and symptom burden as well as reduced healthcare utilisation [10]. In pancreatic cancer specifically, early referral to PC has been associated with reduced use of futile and aggressive anti-cancer treatment towards the end of life [11], fewer acute hospital admissions and reduced healthcare costs [12]. Additionally, PC interventions have been associated with reduced depression and stress in carers of patients with advanced cancer [13].

According to an international consensus recommendation, referrals to specialist PC should be made within 3 months of advanced cancer diagnosis for patients with a median survival of one year or less [14]. Consequently, in line with this recommendation, all non-resectable patients with pancreatic cancer should ideally be referred to PC. However, a cross-sectional study conducted by Beesley et al. in a cohort of patients with pancreatic and ampullary cancers found that only 59% of non-resectable and 27% of resectable patients accessed PC services [15].

To date, there are limited studies which have explored characteristics associated with use of PC specifically in pancreatic cancer and those that have been reported are largely based on the American, Medicare-enrolled population [16]. Studies which have investigated PC use in the Australian context have tended to do so through the limited lens of in-patient PC admissions, chemotherapy in the last 30 days of life and place of death [17]. Consequently, this study aims to describe the characteristics of patients with pancreatic cancer who present to a specialist PC service in Victoria, Australia, using a comprehensive PC admissions dataset, to inform our understanding and guide potential opportunities to promote equitable access to PC. Given the variability in the cancer care management approaches across Australian States and Territories (for example, New South Wales employs a centralised model of care for patients with pancreatic cancer), this study focussed only on Victoria as a case study. Specifically, this study aims to understand the association between patient characteristics and symptom burden at initial presentation to a Victorian specialist PC service, as well as compare the demographic characteristics of patients referred to PC to the overall Victorian population of patients with pancreatic cancer.

Between 2014 and 2020, 2890 eligible first episodes of specialist PC were identified. The majority of episodes had patients recorded as being male (52.5%), born in a country other than Australia (53.1%), English speaking (84.4%), residing in major cities (89.2%) and aged ≥ 65 years (77.5%). Public hospitals were the predominant referral source (50.5%) (Table 1).

This study presents one of the largest series of symptom and problem profiling in patients with pancreatic cancer at presentation to specialist PC. Our findings indicate that the likelihood of experiencing high symptom distress at the commencement of first PC episode decreased over time, perhaps suggesting that timing of access to PC has improved over time. This is also coupled with the finding that general access to PC has increased over time. However, given the majority of first PC episodes began at a deteriorating phase, coupled with the fact that 32% of episodes ended in death, this may suggest that access is not occurring early enough. Aside from symptom relief, early referrals to PC also offer the opportunity for early discussion of goals and setting plans which may mitigate escalation of care at the end of life. Consequently, implementation of early PC in routine practice, as being explored by the Care Plus trial [24], may be highly beneficial for both improving patient outcomes and reducing acute hospitalisations.

Compared to the state-wide pancreatic cancer population, a significantly smaller proportion of the PCOC cohort were registered as residing in a regional or rural area of Victoria. This may be explained by the fact that regional and rural areas often have different models of PC (for example, each health region/district may not have a dedicated specialist PC team and may instead rely on a range of staff who work across multiple services). These services are not defined as ‘specialist’ PC care (and as such, are not included in the PCOC dataset). Surprisingly, symptom distress at initial presentation to PC for patients residing in regional and rural areas was comparable to those residing in metropolitan regions. However, given the limited capture of this sub-population, it is important to further interrogate symptom distress more broadly across all patients with pancreatic cancer who reside in regional and rural areas, particularly for those who access generalist supportive care only. Equitable access to high-quality specialist PC in regional areas may be mediated through a hybrid model of care that includes both in-person and telehealth consultations, with trained specialist PC providers [25].

High distress from fatigue was the most commonly experienced issue, affecting more than 30% of our cohort, followed by appetite distress (affecting 21% of the cohort). These findings are unsurprising as previous studies have indicated that fatigue- as well as appetite-related distress are commonly experienced issues in the pancreatic cancer population [26]. Additionally, our findings also mirror previous studies in that pancreatic cancer patients were more likely to experience high fatigue- and appetite-related issues compared to patients with other lower and higher survival cancers [27]. Facilitation of early access to multidisciplinary supportive care, particularly to dietetic support, is highly desired by patients with pancreatic cancer [28] and should form a key priority area in the delivery of care.

Surprisingly, only ~ 20% of patients experienced high distress related to pain at presentation to PC. A recent study exploring the symptom burden of advanced pancreatic cancer patients found that the median score at baseline for pain was 1, when measured along the ESAS [3]. Our findings also mirror the distribution of pain distress scores reported by other studies using PCOC data [19]. However, literature also suggests that pain affects 80% of patients with pancreatic cancer, with nearly half requiring strong opioids for pain relief [29]. It is also possible that many patients included in our study had received pain management support prior to their first episode of PC; however, this data is not collected through PCOC. The true prevalence of distress related to pain in pancreatic cancer, along the disease trajectory, requires further interrogation so that patients can be offered appropriate support when most needed.

With the exception of pain, admission to an in-patient PC service was generally a predictor of lower symptom distress/problem severity at initial presentation, compared to community setting. A study conducted in a US-based community-dwelling population of patients with advanced cancer reported similar findings in that a significant proportion of patients reported moderate/severe symptom distress when first engaging community PC [30]. Qualitative studies involving patients and caregivers as well as community PC service providers may help to identify barriers and enablers of timely access to community-based specialist PC.

Fatigue at initial presentation was lower in pancreatic cancer patients who preferred a language other than English, with no differences observed in pain, appetite or psychological distress, despite a higher proportion presenting when unstable, deteriorating or terminal. Whilst certain cultural lifestyle factors may potentially enable better symptom management and therefore explain the reduced likelihood to experience fatigue, it is difficult to elucidate the true causation as literature regarding symptom distress in culturally and linguistically diverse (CALD) cancer patients is limited and conflicting. For example, one study found that migrants with cancer were more likely to report worse health-related QoL and higher emotional distress compared to Anglo-Saxon patients [31], whilst another found that CALD patients reported a similar level of symptom distress (measured through the Distress Thermometer and Problem Checklist) to non-CALD patients [32]. Similar to our findings, a recent study using PCOC data reported that cancer patients preferring non-English languages were less likely to report symptoms or problems, except for family/carer problems [33].

Language has been identified as a major barrier to the assessment and management of symptoms in PC [34]. System-level barriers, such as structural racism, may also hinder accurate reporting or disclosure of symptom distress [35]. One way to mitigate these issues is to ensure that Non-English-speaking patients are provided with opportunities to participate in research studies exploring health service delivery assessment [36]. Additionally, tools which are validated in languages other than English should be routinely used when required. It remains unclear whether translated versions were used with patients represented in this dataset. Finally, it is important to integrate PC into standard cancer care pathways so that it is considered at key points along the care continuum [37]. Our finding that a significantly greater proportion of patients who preferred a language other than English were deceased at the end of their first episode of PC suggests that timely access ought to be specifically explored in this population.

Family/carer problem scores were significantly higher at presentation for patients with pancreatic cancer who preferred a language other than English, compared to English-speaking patients. Interestingly, whilst this effect was observed within families/carers, it was not reflected in the psychological distress scores of patients themselves. A recent qualitative study into the supportive care needs of patients diagnosed with pancreatic and oesophagogastric cancers has reported that patients often rely on carers for emotional support and do not feel the need to seek professional support. Their carers, however, desire professional support for themselves [28]. Whilst the unmet needs of cancer carers, particularly those who are of a CALD background, have received limited attention to date, a survey conducted by Carers NSW found that carers from CALD backgrounds report higher levels of psychological distress and less social support [38]. Further investigation of the unmet needs of CALD cancer carers is required to inform, design and implement culturally appropriate services and resources to reduce the burden they may experience.

Although the odds of presenting with high pain, psychological/spiritual distress or family/carer problems decreased with age, it is important to interpret these findings with caution as the reason behind this association is not clearly understood. It has been posited that elderly patients are generally more emotionally resilient, have fewer competing demands and are more economically stable compared to younger patients [39, 40] thereby potentially explaining our findings. However, elderly patients may also under-report distress due to various factors, such as memory/hearing loss, confusion, fear of being a burden and/or stoicism [41].

To our knowledge, this is the first in-depth exploration of the characteristics and symptom burden of patients with pancreatic cancer who present to specialist PC. Our study, however, has several limitations. Despite capturing 90% of specialist PC services across Australia, PCOC is a voluntary service and therefore does not capture all PC presentations, thereby limiting the generalisability of our findings to the overall pancreatic cancer population. Additionally, given our inability to uniquely identify patients across all services, it is possible that some patients who presented to multiple different services during the eligible time period may have been counted more than once. Lack of clinical data and date of diagnosis did not allow us to look at symptom distress along the disease trajectory, timeliness of access to PC, nor survival outcomes. Socioeconomic status for patients was also not available. Future studies should interrogate the association between SES and PC use as low SES has been shown to be a predictor of poorer access to healthcare services [16]. Nevertheless, to our knowledge this is the largest series of symptoms and problem profile of people with pancreatic cancer and represents an important baseline for future studies in this uncommon cancer.

Although a large proportion of first-time PC episodes in pancreatic cancer begin when the patient is deteriorating and end in death, high symptom distress (with the exception of fatigue) at presentation is relatively low and not generally associated with a patient’s demographic characteristics. However, the higher likelihood of presenting to community-based specialist PC with high symptom distress suggests earlier referral pathways require further interrogation. Access to PC ought to be explored in CALD communities given a greater proportion of deaths amongst non-English-speaking patients at initial presentation and higher family/carer problems scores were observed. Additionally, reasons for underrepresentation of patients residing in regional and remote areas who access specialist PC requires further investigation.