Is an intensive treatment regimen for newly diagnosed type 2 diabetes patients cost-effective? Economic evaluation of the 5-year results of the ADDITION study.

A new paper in Diabetic Medicine reports an economic evaluation of the ADDITION study, based on the five year follow-up data.  The Anglo-Danish-Dutch study of Intensive Treatment In peOple with screeN detected diabetes (ADDITION) is a prospective randomised controlled trial of screening and intensive treatment of newly diagnosed type-2 diabetes patients.

The five-year outcomes, published in the Lancet, observed a small, non-significant reduction in the incidence of cardio-vascular events and death associated with an intensive treatment regimen on diagnosis of type 2 diabetes.

However, in the newly published economic evaluation (based on the UK data), this translated into virtually no difference in overall health related quality of life (measured in QALYs gained) over the five years.  On the cost side, the extra cost of the intervention was not offset by any reductions in treatment costs for acute events, and so the intervention appears not to be cost-effective.

But this is not the whole story.

The economic evaluation also projects the five-year costs and outcomes to a longer time horizon (30 years), using the UK Prospective Diabetes Study Outcomes Model.

Over the longer time horizon, the full benefits of those averted events and fatalities begin to reveal themselves in the added QALYs with an accrued benefit of 0.0465 QALYs per patient over 30 years (significant at 95% level).  However, the accrued extra cost means that the incremental cost-effectiveness ratio is estimated at £37,500, still above the upper limit of £30,000 as used by NICE.  So even over 30 years, the intensive treatment regimen may not be cost-effective.

But again, this may not be the complete story.

Our results found an improvement in cost-effectiveness as the time horizon increased, although this seems to bottom out at 20 years.  Exactly why this is worthy of investigation, but £37,500 per QALY is not hugely in excess of NICE thresholds, and there is a 30% probability that the ICER is actually below £30,000.

We concluded that the intervention could only be cost effective if it could be delivered for around 2/3rds the cost.  However, it may be worth looking in more detail about some of the assumptions in the analysis.

Any decision model based analysis relies on many assumptions over the choice of model structure, health state valuations and so on.  It is considered good practice to be conservative in those assumptions.  That is, given a choice of alternatives with no rationale for preferring one over another, it is better to choose that which favours control.  That way, if your results suggest the intervention is cost-effective, this will be despite ‘loading the dice’ against the intervention.

In this case we assumed perfect adherence by GPs to the intensive regimen (for which they were reimbursed).  This included a certain number of contacts per year as well as an expected number of prescriptions etc.  If it turns out GPs and their patients are mere humans, then they may not have adhered perfectly to the treatment algorithms.  We may therefore have overestimated the cost of the intervention arm.

Secondly, recruitment to the ADDITION study took place between 2001-06.  Since then there have been several sets of guidance recommending more intensive treatments for all patients diagnosed with type 2 diabetes.  Therefore the routine treatment received by our control arm is likely to have intensified over time, such that it became closer to the intensive treatment arm.  We may therefore have underestimated cost in the control arm, and underestimated incremental outcomes.  The net effect is that our analysis may have over-stated the incremental cost-effectiveness ratio.  If this had still been comfortably within the bounds of acceptability (i.e. below £20,000) we would have had a nice positive story to report.  However this was not to be.  Thus are the perils of any pragmatic randomised controlled trial with long term follow-up.

In order to address these issues, we are now looking in more detail at the treatment patterns received by a sample of patients in the intervention and control arms, by extracting data from their primary care records.  With this new information we will firstly be able to see how good the GPs were at getting their patients to adhere to the intensive treatment protocol, secondly to assess whether the control arm did experience any ‘intensity drift’ in their treatment over time, and finally we will be able to revisit the cost-effectiveness analysis to see if we should change our conclusions.

The fat lady hasn’t yet begun her final aria.

This entry was posted in Blog and tagged , , , , . Group: . Bookmark the permalink. Both comments and trackbacks are currently closed.
  • The Cambridge Centre for Health Services Research (CCHSR) is a thriving collaboration between the University of Cambridge and RAND Europe. We aim to inform health policy and practice by conducting research and evaluation studies of organisation and delivery of healthcare, including safety, effectiveness, efficiency and patient experience.