This study identified that, according to expert assessment, only 14 (6.7%) of 208 inpatient admissions amongst patients with palliative care needs in two acute settings in England were ‘potentially avoidable’. Those patients whose admissions were considered to be ‘potentially avoidable’ had a median age of 84. Half of the patients lived in residential or nursing care, and we concluded that most could have received care in this setting. The proportion of ‘potentially avoidable’ admissions identified is much lower than that suggested by previous research [10, 11].Several reasons for this can be identified.
Firstly, and most importantly, admissions were considered by palliative medicine consultants from the standpoint of the admitting clinician and within the context of existing local health and social care services. In contrast, previous studies defined ‘appropriateness’ in relation to an ideal configuration of end of life care services, rather than the less than ideal situation that exists at the moment. A key strength of the study is the involvement of experienced local palliative medicine consultants in determining admission appropriateness who, crucially, have very good knowledge of the local configuration and extent of health and social care services for this patient group.
Secondly, their decision-making was also informed by the wider context of clinical risk and uncertainty that is critical to how these decisions are made in real-life acute situations. Two –thirds of the sample was admitted to hospital at a time when routine services might not be available. In contrast to previous studies [10, 11], we also included a sample who were alive at the time of initial data collection and who met recognised diagnostic and prognostic criteria for palliative care need. We were also, therefore, able to include patients with palliative care needs who did not die during that hospital admission; this is important given the high number of hospital admissions many of these patients experience during the last year of life.
Whilst the absolute numbers of the group of patients classified as experiencing a ‘potentially avoidable’ hospitalisation was small, their socio-demographic and clinical characteristics do provide an interesting insight into where efforts to reduce potentially avoidable admissions would be most fruitfully expended. Most were older (median age 84 years) and lived in nursing or residential care. Crisis admission to hospital of nursing home residents remains an issue, with up to 67% of hospitalisations from nursing homes identified as avoidable . There is evidence that older people in residential care settings receive variable quality end of life care, because of clinical and organisational factors . A recent study which aimed to identify key factors influencing end of life care in nursing homes identified significant difficulties in implementing appropriate end of life care due to factors including lack of support from other agencies, lack of out of hours support, lack of access to training, variable support by general practitioners (GPs), and reluctance among GPs to prescribe appropriate medication . In addition, high continuity of primary care has been shown to be important in reducing potentially avoidable hospitalizations amongst older people . Whilst good Advance Care Planning (ACP) processes have the potential to help prevent admissions, where this is in line with the patient’s wishes, ACP remains under-developed within a UK context .
The majority of patients in this study accessed the hospital following presentation to the Accident & Emergency Department and were admitted out of hours. In the UK unplanned admissions have risen steadily over the past 10 years. Most of these ‘unplanned’ admissions occur ‘out of hours’, and most are admissions via A&E . The King’s Fund Report (‘Avoiding hospital admissions’) found that many admissions labelled as ‘avoidable’ or ‘potentially avoidable’ are necessary as the only place that the services which are clinically required are available is in the acute sector . The report found that certain interventions could reduce inpatient admissions; for example an emergency medicine consultant reviewing patients on admission could reduce inpatient admissions by 12% and specifically reduce admissions to the medical assessment unit by 21%. Whilst this evidence was generated from a general hospital population rather than from patients with palliative care needs, it is reasonable to assume that interventions of this nature would have a similar impact within a palliative care population. However, we are very far from being in the situation where a senior clinician is available to review every hospital admission and, again, many of these admissions could only be prevented if patients had good access to a comprehensive range of primary and community services.
There are a number of recognised limitations to this study. Assessments of potentially avoidable admissions were made by Palliative Medicine consultants from the standpoint of the admitting clinician, using expert judgements based on complete case note information. This subjective approach may be open to individual bias, however in the absence of any validated protocols to identify potentially avoidable admissions [17, 18], context specific expert judgement was deemed the most appropriate way to classify admissions.
Seventy four patients were excluded as complete data were not available. Whilst there were no significant differences in age and gender between these patients and those who were included, other differences between the groups may exist. Therefore the generalisability of the dataset should be considered with caution.
Statistical comparisons between appropriate admissions and potentially avoidable admissions were not possible due to the small numbers of potentially avoidable admissions. Future research should seek to address this gap in current understanding in order to tailor interventions to those most at risk of potentially avoidable admissions.