Why are they linked?
In an earlier briefing we noted that deprivation has a significant effect on 2005/06 reference costs. The trend is also observable in the latest reference cost data for 2006/07 (top chart).
We said that there are two broad schools of thought. It may simply be that having more money leads to higher reference costs – deprived areas have long since had more money allocated to them because of greater ill-health. On the other hand, patients in deprived areas may be more costly to treat. For example, their length of stay, which is a prime cost driver, may be necessarily longer. But the reference cost index (RCI) is already adjusted for casemix, so any variation would have to be within individual narrowly specified conditions, ie individual health resource groups (HRGs). We think we can now answer that question.
Links between length of stay and deprivation are hard to spot. Out of the top 50 HRGs, accounting for 4.9 million episodes per year, we found only one (F47, general abdominal disorders >70 w/o cc, representing only 0.14 million episodes) which exhibited any significant link. Overall, we found no link between length of stay (LOS) (compared with that expected for casemix in the top 50 HRGs) and deprivation, as the middle chart illustrates.
We also looked at hospital mortality (rates per episode). That too revealed no significant relationship (bottom chart). There are actually slightly fewer deaths per episode in more deprived areas (though this is not quite statistically significant) – possibly due to the lower average age of more deprived populations. Our analysis was standardised for casemix, though not for age, but as we found no residual variation of mortality with age we conclude that any age effect is essentially fully factored into the HRG casemix variation.
On the basis of this evidence, the answer to why trusts in deprived areas have higher reference costs is simply that PCTs in deprived areas have more money (80 per cent of the variation in allocations per head is linked to deprivation) and have historically paid their trusts accordingly.