Health economic assessment of medical technology in chronic progressive diseases : multiple sclerosis and rheumatoid arthritis

Abstract

This thesis developed a method for economic evaluation in non-fatal chronic progressive diseases, applied to multiple sclerosis (MS) and rheumatoid arthritis (RA). In economic evaluation different treatment strategies are compared in terms of their costs and their health benefits. If an alternative has a better outcome, but is also more costly, as is the case for the new treatments introduced for MS and RA, the extra cost for the incremental benefit, i.e. cost-effectiveness, is estimated.

The challenge for economic evaluation in chronic progressive disease is the complexity of these pathologies where acute symptoms are present permanently or intermittently, while at the same time the severity of the disease progresses. Symptoms can or cannot represent risk factors. Treatment can be symptomatic or disease modifying or both, and economic evaluation must therefore take both symptoms and progression. into account.

The method developed describes disease progression with functional disability measures, relates costs and quality of life (utility) to different levels of functional disability, incorporates temporary symptoms, and expresses the effectiveness of treatments as the gain in quality-adjusted life-years (QALYs). Disease progression is estimated from epidemiological cohort studies, costs and utility are collected in specially designed large observational studies, the effectiveness of treatment is taken from clinical trial data and the cost-effectiveness is estimated using long-term models (Markov models).

The studies show that the models capture disease progression and symptoms, costs and quality of life adequately, but that the results are highly dependent on the quality of the underlying data and the assumptions made for extrapolating to the long term.

The three papers in MS illustrate how a model can be developed from a clinical trial, improved over time, as new date become available, and adapted to different patient populations. Extrapolation of the clinical trial data underestimated disease progression compared to the natural history data. As a consequence, the cost per QALY gained with 3-year treatment over a 10-year timeframe dropped from 342'700 SEK in the first study to 257'200 SEK in the second. When the model is applied to patients with active disease at higher risk of progression, the cost per QALY gained was further reduced to 72'000 SEK.

The three papers in RA illustrate how a model can be based on epidemiological data and adjusted to allow incorporating clinical trial data. The models based on two independent epidemiological cohorts of newly diagnosed patients in Sweden and the UK showed a strikingly similar outcome, with patients having an average of 5.1 QALYs over 10 years. However, costs were very different, due to large differences in indirect costs between the two countries. The models were adapted to patients with more advanced disease treated with infliximab in a one-year trial, and a loss of effectiveness after the trial was incorporated. Treatment was shown to be cost-saving in Sweden, while the cost per QALY gained in the UK was 18'400£. It was not possible to adapt the model to patients with very long standing disease treated in clinical practice with etanercept or infliximab due to lack of data in the epidemiological data sets. Treatment was therefore compared to baseline, resulting in a cost per QALY gained between 340'000 SEK and 500'000 SEK.

The thesis shows that the method is appropriate and feasible, but the results are highly dependent on the availability of appropriate data.