9.1 Introduction

There has been much debate and discussion on terminal care in the last year. This has culminated with the White Paper “Our Health Our Care Our Say” (Department of Health 2006) and many national and regional documents which expound the need for a good death. This is often discussed in relation to cancer care but not often discussed in relation to non–cancer care. Cancers are increasingly becoming a chronic illness and over 50% of cancer patients do not die.

There has also been a considerable amount of work undertaken in the region on the role of end of life services when looking after patients with long-term conditions. This is essential in preventing the admission of patients into acute hospitals when they are dying. When patients are asked about their preferred place of death – only about 10% want to die in hospital and about 70% want to die at home. The remainder want to die in hospices.

This chapter aims to compare the place of death for different PCTs in West Midlands (WM) Region. This is undertaken by scrutiny of place of death codes found in the ONS death certificate data. In addition, analysis is presented to focus the debate on long term conditions and admission prevention.

9.2: Background

ONS Death Certificates contain a variety of data fields. These include place of residence and place of death. The place of death uses the NHS provider lookup file for known providers of healthcare. There are many thousands of providers – including hospitals, hospices, nursing homes, care homes, prisons, psychiatric hospital, etc. The numerous providers are also categorised using an ONS classification known as the establishment type. This is a list of codes and are described as follows:

ONS standard codes for establishment type as used on ONS death certificates.

The codes have been aggregated for the purposes of this chapter to:

• Hospital
• Care Home
• Home
• Hospice
• Other

For an in depth breakdown of the codes please refer to the data disc which accompanies this publication.

9.2.1: ONS Death Certificate Codes

ONS records diagnoses associated with deceased individuals on a series of data fields. There are 3 diagnosis fields used; primary (for the main cause of death), secondary (for the underlying cause of death) and tertiary (for ancillary conditions not necessarily associated with death). The data analysed below relates to deaths during 2002 – 2003.

The analyses presented in this chapter are based on the following codes being used in the primary and secondary diagnostic fields only

Table 9.1: Diagnostic labels and ICD10 codes

The diagnosis data presented is not mutually exclusive – there are some individuals who have chronic pulmonary disease and heart failure. However the different places of death are the most important facet of this analysis.

9.3: Methods

Deaths were aggregated using the diagnostic and establishment labels described above. In addition, geographical information was attached to the data to provide breakdowns by PCT.

9.4: Results

The following tables and figures provide information relating to the percentage of deaths with death certificates coded by PCT. This is a proportionate mortality comparison and is not age sex standardised.

9.4.1: All Deaths

Just over half of all deaths occurred in hospital. Twenty one percent of deaths occurred at home – in stark contrast to the stated wish that about 70% of people expressed a preference to die at home. In addition to this, different clinical categories have different percentages of sufferers dying in other locations. Almost 60% of dementia patients die in care homes and most of the remaining die in hospital. Twenty five percent of cancer patients die at home – the highest percentage when compared to heart failure and respiratory failure. This is probably because of the palliative care services that are able to support patients and also because of the gradual disease trajectory these patients have. In contrast the vast majority of chronic renal failure patients die in hospital (84%). The presence of care homes and hospices can be seen in non-cancer and cancer related deaths.

Table 9.2: Percentage of deaths in each establishmen

The following tables and figures provide an indication of the relative percentages of deaths by condition by PCTs. Deaths in other places are so small, (less than 1% in total) and are not presented here. Percentages are presented here but absolute numbers for each of the conditions are presented on the data CD which accompanies this publication.

9.4.2: Cancer deaths

The percentage of cancer deaths that take place in hospital vary enormously. The PCTs with the highest cancer deaths in hospital tend to be the PCTs with the highest deprivation levels.

Table 9.3: Cancer deaths by PCT and establishment type

There is an inverse relationship between hospital deaths and hospice deaths. Each dot in the following series of graphs represents a PCT :

Figure 9.1: Cancer Deaths  - Hospice vs Hospital, WM PCTs ONS Death Data

Figure 9.01: Cancer Deaths - Hospice vs Hospital, WM PCTs ONS Death Data

There is also a slight negative association between hospice deaths and home deaths.

Figure 9.2: Cancer Deaths -  Hospice vs Home,  WM PCTs ONS Death Data

Figure 9.02: Cancer Deaths - Hospice vs Home, WM PCTs ONS Death Data

These data suggest that PCTs with hospice provision experience fewer hospital deaths and home deaths from cancer. This sounds intuitively correct as the presence of hospices enable terminally ill patients to have an alternative place to be admitted when they are dying. As a commissioning model, local hospice provision is beneficial as it prevents admissions to hospital. However close scrutiny of the relationship between home deaths and hospice deaths shows a similar if less significant pattern where PCTs with hospices have fewer home deaths.  This is not necessarily what the majority of terminally ill patients want.

9.4.3: Pulmonary Deaths

The majority of deaths occur in hospital and there is a large variation across different PCTs.

Table 9.04: Pulmonary deaths by PCT and establishment type

In a similar comparison to hospital deaths vs hospice deaths, it is apparent that those PCTs which have pro-rata fewer hospital deaths for pulmonary disease have more deaths in care homes and at home. It might be that this association is causative - ie those PCTs with greater provision of continuing care and social service provision have more of their elderly frail population in care homes (Figure 9.03) or domiciliary care (Figure 9.04), some of whom have chronic pulmonary disease.

Figure 9.3: Place of Death for Pulmonary Disease - WM PCTs – Hospitals vs Care home

Figure 9.03: Place of Death for Pulmonary Disease - WM PCTs - Hospitals vs Care Homes

Figure 9.4: Place of Death for Pulmonary Disease - WM PCTs – Hospitals vs Home

Figure 9.04: Place of Death for Pulmonary disease - WM PCTs - Hospitals vs Home

9.4.4: Heart Failure Deaths

The majority of deaths occur in hospital and there is a large variation across different PCTs.

Table 9.5: Heart failure deaths by PCT and establishment type

In a similar comparison to hospital deaths vs hospice deaths, it is apparent that those PCTs which have fewer deaths in care homes and at home have more deaths in hospital. When deaths in hospital are compared to deaths at home no relationship (or a very weak one) in observed. (Figure 9.6).

Figure 9.5: Place of Death for Heart Failure - WM PCTs  - Hospital vs Care Home

Figure 9.05: Place of Death for Heart Failure - WMPCTs - Hospital vs Care Home

Figure 9.6: Place of Death for Heart Failure - WM PCTs – Hospital vs Home

Figure 9.06: Place of Death for Heart Failure - WM PCTs - Hospital vs Home

9.4.5: Renal Deaths

Unlike all deaths described above patients with renal failure almost always die in hospitals. The absolute numbers of deaths are very low and the small number of renal patients who do die, invariably are admitted as emergency admissions.

Table 9.6: Renal deaths by PCT and establishment type

9.4.6: Dementia Deaths

Table 9.7: Dementia deaths by PCT and establishment type

As can be observed many patients die in care homes. There appears to be no relationship between deaths in hospital and deaths at home (Figure 9.08).

Figure 9.7: Place of Death for Dementia - WM PCTs – Hospital vs Care Home

Figure 9.07: Place of Death for Dementia - WM PCTs - Hospital vs Care Home

Figure 9.8: Place of Death for Dementia - WM PCTs – Hospital vs Home

Figure 9.08: Place of Death for Dementia - WM PCTs - Hospital vs Home

No relationship can be observed between death in hospital and death at home.

9.5: Discussion

The results provide a useful tool for reflection on commissioning in different areas.

Commissioning is currently focused on steering service provision away from hospitals. This is because of a variety of reasons including the new white paper “Our Health, our Care, our Say” as well as fiscal and primary based commissioning models. The different patterns of deaths give some indication of where the priorities lie in relation to long term conditions. The other negative opportunity costs are that the long term condition patients are more costly because they aggregate to HRGs with the “complication and co-morbidities” categories or the complex elderly HRGs (with the suffix code 99). All of these are more expensive than the simple HRG codes. This is further exacerbated by the ageing population (see Chapter 6) which also increases the number of old age related HRGs and thereby PbR costs

Community based provision can either be social services funded or voluntary / hospice funded.

Social services provision appears to be closely associated with patterns of location of deaths for specific conditions. This is so that those areas where there appear to be fewer deaths in care homes have more deaths in hospital. This might be cause, effect or merely an association. In the above sections the suggestion has been made that the provision of alternative long term beds might be the cause of reduced hospital admissions. There is a contrary view that the higher proportion of deaths are present in hospitals because of delayed discharge and long stays.

To consider this further, the same patterns are present with hospice provision. On this basis, it might seem that this relationship is in fact causative. This is because where there is a hospice, more cancer patients die in the hospice out of choice and availability. This can be seen in the graphs above in that PCTs where a hospice is present also have fewer deaths in hospital.

Renal failure patients completely buck the trend as there is a paucity of community placements where renal patients can be cared for. Perhaps this is where the greatest influence can be had in terms of alternative places of care.

Inappropriate admissions are currently a focus for commissioning opportunities and in particular payment by result. Deaths in various settings give an indication of where deaths occur and give an indication of where the best interventions are focused. For example PCTs which have relative under provision of care homes might benefit from bolstered outreach services for heart failure and pulmonary disease.

The dearth of hospices is associated with higher emergency admissions to hospitals for patients with cancer. The supreme irony of hospice death data is that the most deprived populations with the highest cancer incidence and death rates have poorest access to palliative care places in hospices. It is probably the most extreme form of inverse care law.

Reference:

• Our Health, Our Care, Our Say: A new direction for community services, Department of Health, 30 January 2006, ISBN: 0101673728