The research recommendations from this programme fall into three areas: (1) research recommendations about the measures themselves, (2) research recommendations on the models of palliative care and (3) research recommendations about the casemix classification.
Research recommendations about the measures
Implementation research using the IPOS – particularly on the best ways to implement it into clinical services and how to assess which is the best version for use (patient-reported or proxy-reported) – is much needed. Some of this research has been subsequently undertaken105 and more research is already in process (see Bradshaw et al.106 and www.hyms.ac.uk/research/research-centres-and-groups/wolfson/resolve).
In addition, the IPOS needs to be adapted for specific advanced disease populations, for example those with end-stage kidney disease107 and heart failure108 – this work is developing rapidly.
Research recommendations about the models of palliative care
Our theoretical framework to understand the complexity of palliative care needs is a useful advance for both clinical practice and research, but we need to study more fully certain aspects of complexity within the exo-system (services/systems) and macro-system (societal influences). These areas are much less studied than the micro- and meso-levels, and yet they are crucial if integrated care (i.e. care that cuts across service and community boundaries) is to become a reality. It is important to determine and explore socioeconomic status, and how this affects palliative care needs. ‘Cumulative’ and ‘invisible’ complexity may interact with socioeconomic status and so influence some of the known inequities in provision according to socioeconomic position.109
The research on models of care needs to be extended and externally validated beyond the C-CHANGE sites, which – by nature of their willingness to participate – may not reflect all models of specialist palliative care.
Although a systematic review of the components of models of palliative care has been undertaken,110 it is not yet clear which components are used in different models of care, let alone whether or not they are clinically effective and cost-effective. Brereton et al.111 identified the importance of defining and describing the components of models of care to differentiate them in both practice and research, and to truly understand clinical effectiveness and cost-effectiveness. We have progressed work on defining models of care (see Workstream 2: models of specialist palliative care),81 but much more is needed. There is also a major need to understand how primary and community services work with specialist palliative care; some of our data suggested that specialist palliative care models were significantly adapted according to the primary- and community-based services for end-of-life care available, and especially in relation to COVID-19.
Research recommendations about the casemix classification
The casemix classification for specialist palliative care provides a standard way to measure complexity of needs, enable services to compare workload between teams and determine whether or not outcomes are achieved as expected for different levels of complexity. The classification could also – if used judiciously – allow teams and services to make a better case for sufficient resources, and – in the longer term – support casemix-adjusted outcome measurement. There are a range of implications for research.
First, these casemix classes should be applied within research studies to better characterise study populations and therefore improve understanding of generalisability of research findings.
Second, whether or not these casemix classes work across the full range of non-cancer conditions should be tested, especially as the proportion of people with multiple long-term conditions increases.
Third, further research should determine how these casemix classes hold over time as the population age distribution changes and the patterns of end-of-life conditions changes. The classes developed within this programme are similar to those developed in the UK by the Palliative Care Funding Pilots58 and to the Australian casemix classification,57 with some of the same criteria found to predict variance in resource use, notably symptoms, palliative Phase of Illness, functional status and family distress. The combination of criteria within each class is, however, somewhat different in this programme than in these earlier studies. These differences may reflect a changing palliative care population since the Australian classification and UK Funding Pilots were completed. The proportion of those with non-cancer conditions receiving specialist palliative care, for instance, is increasing steadily112 and 20% of participants recruited had non-cancer conditions to reflect this change.
Fourth, it is important that these casemix classes and related costs are tested across a wider range of NHS and non-NHS services. This study provides reasonably current patient-level cost data for episodes of specialist palliative care. NHS unit costs for specialist palliative care do exist,113 but their reliability is unclear and they do not reflect non-NHS services (which includes most specialist palliative care in the UK). A recent systematic review of the cost of UK palliative care114 found very limited cost evidence (only 10 studies over 20 years), with most studies combining estimates of resource use with potentially unreliable unit cost data. Compared with the 2018 NHS unit costs collected in 2017113 (i.e. at the same time as the collection of cost data for this study), we found actual costs of specialist palliative care – at least in hospital and community settings – to be lower: hospital advisory care cost an average of £73 per day (compared with £201 in NHS unit costs113), and community care was £36 per day (compared with the unit cost for a specialist nurse of £64113). In part, this may reflect the challenges of collecting accurate cost data, particularly in community settings, but it may also reflect the range of staff providing care and the frequency of visits (i.e. less visits under cost constraints leading to reduced costs per day). Models of specialist palliative care are also increasingly diverse,81 especially in the community,110 and workforce shortages are leading to new models, for instance with senior staff supporting less experienced staff with skilled work, a reduced frequency of visits and new models for older people with multimorbidity.100 By contrast, we found that at an average of £716 per day, inpatient palliative care was notably more costly than the 2018 unit cost of £404 per day;113 this may reflect the increasingly complex needs of those needing inpatient specialist palliative care.