Abstract
1. This report places the documented growth in emergency inpatient hospital care in
England in context, by considering how it compares with the experience in Scotland.
2. Scotland is an interesting and important comparator because whilst it has
experienced a very similar pattern of expenditure growth on its NHS, it has adopted a
different policy framework. It thus potentially permits the identification of the role of
policy reform in determining any differential outcome.
3. A first step of our investigation was to examine the comparability of hospital
inpatient activity measures in Scotland and England. We conclude that care is
needed due to terminological differences in the two countries. Some existing studies
have avoided this issue by comparing a narrow definition of emergency care –
admission episodes.
4. We devise a strategy for broadening the measure of activity to include all episodes
of care that follow directly on from an emergency admission. In the case of Scotland
our approach entails a degree of approximation and in the future we hope to test the
sensitivity of our findings to this issue. For the purposes of our comparative study we
find that the results from the narrow definition of emergency care are very similar to
those for the wider definition.
5. Existing studies have focused on a count of episodes but this measure does not
account for the intensity of resource use that is entailed. We therefore consider both
episodes and total bed days. This distinction is both conceptually and practically
important. Whereas episodes have displayed almost incessant growth in per capita
terms, bed days have fallen.
6. In comparing England and Scotland it is necessary to take account of their
different populations, differences in the changing structure of those populations, the
different mixture of treatments that are provided in each country, and differences in
the policies that have been adopted.
7. We adopt a Fixed Effects regression framework that controls for case mix and we
deflate activity by population specific measures in order to control for population
structure and changes. In regard to this second issue we find little evidence of
important differences in the two countries, so that more sophisticated controls for
population have no material effect on the results.
8. Whereas the uncorrected data for episodes show much more rapid expansion in
England than Scotland some of this is explained by case mix. Accounting for case
mix variation also gives a different interpretation of the time path for bed days –
rather than strong negative growth, the data is better explained in terms of near
constancy over time punctuated by downward shifts coincidental with policy
interventions.
9. In regard to emergency episodes we find that England has experienced faster
growth than Scotland, but that the difference is more modest than raw data appears
to suggest and the difference in trend continues to be offset by the lower propensity
for emergency episodes in England relative to Scotland.
10. The key policy interventions we have considered are those associated with PbR
and overall we find that these interventions have moderated the overall expansion of
episodes, although we caution against over-interpreting these results since there is a
complex combination of policy impact effects and effects of changes in the payment
mechanism (tariff).
11. In regard to bed days we find that England has experienced a decline over the
period 1998-2011 both absolutely and relative to Scotland. This decline is not a
simple trend process – the impact of policy appears to be more substantial and more
fundamental than with the episodes measure of activity.
12. We consider simple summary measures of elective activity in the two countries
and do not find any prima facie evidence of England having a faster expansion in
emergency activity on account of its lower elective activity relative to Scotland. There
is more elective activity in England than Scotland over the sample period and as
measured by episodes this activity grows faster in England than in Scotland.
13. Our fixed effects regression framework provides a tool for examining and
comparing the growth in emergency activity in Scotland and England on an HRG by
HRG basis. We present some examples and recommend that a further analysis on
this basis is a potentially fruitful avenue of research.
14. Our key conclusions are:
Care is needed when comparing emergency hospital activity in Scotland and
England.
It is important how activity is measured – episodes of care do not give the
same picture as bed days.
It is important to account for the differences between the two countries in
terms of case-mix, population and policy choices.
After accounting for case-mix, population and policy differences, England has
shown faster growth in emergency episodes but has a persistently lower rate
of those episodes.
Using bed days as a measure of activity, England has had persistently lower
emergency hospital activity than Scotland.
Payment by Results is associated with reductions in emergency hospital
activity in England.
England in context, by considering how it compares with the experience in Scotland.
2. Scotland is an interesting and important comparator because whilst it has
experienced a very similar pattern of expenditure growth on its NHS, it has adopted a
different policy framework. It thus potentially permits the identification of the role of
policy reform in determining any differential outcome.
3. A first step of our investigation was to examine the comparability of hospital
inpatient activity measures in Scotland and England. We conclude that care is
needed due to terminological differences in the two countries. Some existing studies
have avoided this issue by comparing a narrow definition of emergency care –
admission episodes.
4. We devise a strategy for broadening the measure of activity to include all episodes
of care that follow directly on from an emergency admission. In the case of Scotland
our approach entails a degree of approximation and in the future we hope to test the
sensitivity of our findings to this issue. For the purposes of our comparative study we
find that the results from the narrow definition of emergency care are very similar to
those for the wider definition.
5. Existing studies have focused on a count of episodes but this measure does not
account for the intensity of resource use that is entailed. We therefore consider both
episodes and total bed days. This distinction is both conceptually and practically
important. Whereas episodes have displayed almost incessant growth in per capita
terms, bed days have fallen.
6. In comparing England and Scotland it is necessary to take account of their
different populations, differences in the changing structure of those populations, the
different mixture of treatments that are provided in each country, and differences in
the policies that have been adopted.
7. We adopt a Fixed Effects regression framework that controls for case mix and we
deflate activity by population specific measures in order to control for population
structure and changes. In regard to this second issue we find little evidence of
important differences in the two countries, so that more sophisticated controls for
population have no material effect on the results.
8. Whereas the uncorrected data for episodes show much more rapid expansion in
England than Scotland some of this is explained by case mix. Accounting for case
mix variation also gives a different interpretation of the time path for bed days –
rather than strong negative growth, the data is better explained in terms of near
constancy over time punctuated by downward shifts coincidental with policy
interventions.
9. In regard to emergency episodes we find that England has experienced faster
growth than Scotland, but that the difference is more modest than raw data appears
to suggest and the difference in trend continues to be offset by the lower propensity
for emergency episodes in England relative to Scotland.
10. The key policy interventions we have considered are those associated with PbR
and overall we find that these interventions have moderated the overall expansion of
episodes, although we caution against over-interpreting these results since there is a
complex combination of policy impact effects and effects of changes in the payment
mechanism (tariff).
11. In regard to bed days we find that England has experienced a decline over the
period 1998-2011 both absolutely and relative to Scotland. This decline is not a
simple trend process – the impact of policy appears to be more substantial and more
fundamental than with the episodes measure of activity.
12. We consider simple summary measures of elective activity in the two countries
and do not find any prima facie evidence of England having a faster expansion in
emergency activity on account of its lower elective activity relative to Scotland. There
is more elective activity in England than Scotland over the sample period and as
measured by episodes this activity grows faster in England than in Scotland.
13. Our fixed effects regression framework provides a tool for examining and
comparing the growth in emergency activity in Scotland and England on an HRG by
HRG basis. We present some examples and recommend that a further analysis on
this basis is a potentially fruitful avenue of research.
14. Our key conclusions are:
Care is needed when comparing emergency hospital activity in Scotland and
England.
It is important how activity is measured – episodes of care do not give the
same picture as bed days.
It is important to account for the differences between the two countries in
terms of case-mix, population and policy choices.
After accounting for case-mix, population and policy differences, England has
shown faster growth in emergency episodes but has a persistently lower rate
of those episodes.
Using bed days as a measure of activity, England has had persistently lower
emergency hospital activity than Scotland.
Payment by Results is associated with reductions in emergency hospital
activity in England.
| Original language | English |
|---|---|
| Publisher | Centre for Health Service Economics and Organisation |
| Commissioning body | Department of Health (England) |
| Number of pages | 33 |
| Publication status | Published - Feb 2015 |