Discussion

The analysis presented here is based on published estimates of the cost of respiratory diseases in European countries, and on prevalence, incidence and population data as reported in chapter 1 and from WHO and Eurostat publications. There are major gaps in cost estimates in the literature, particularly those of pneumonia/ALRI, as well as the indirect costs of several diseases and all costs of the many respiratory diseases not listed here. Consequently, our estimates are by default, considerable underestimates.

Even with diseases in which good studies are available there is inevitable uncertainty in our overall cost estimates due to the need for extrapolation of data from a small number of countries to all 28 countries of the EU. For inclusion in the present analysis, the profiles of the countries concerned had to share at least some characteristics for the costs to be relevant. This could have an impact on both direct and indirect costs. In asthma and COPD, regression analysis was used to extrapolate from the costs for a few hundred patients, or a registry study of thousands of patients, to a population of some 500 million persons, so errors may grow substantially. However the regression models of cost data give a reasonable fit and no evidence of serious bias. The DALY estimates are published by WHO by region and sub-region and therefore are less prone to error, but there were variations in estimates from different sources; the WHO estimates used here were the lowest estimates available. Even among the limited range of diseases considered, no estimates of indirect costs were available for cystic fibrosis or OSAS. The total estimates presented here must therefore be considered minimum and very conservative.

We have aimed for simplicity of design and methodology and more sophisticated methods might have improved certain aspects. Some costs such as patients’ out-of-pocket costs are missing due to an almost complete lack of data. Furthermore, some indirect costs relating to absence from work and early retirement are missing for several conditions, where there is no relevant literature. On the other hand, the method of costing sickness absence and early retirement by average earnings may somewhat overestimate indirect costs where there is high unemployment.

An alternative approach to estimation of costs is by primary research, such as a multinational study with data collection and costing of each patient using local unit costs, and taking median costs across nations to obtain the overall estimate. Accordini et al. recently carried out such a study of persistent asthma in eleven European countries. They estimated the direct and indirect cost of persistent asthma based on 5 million subjects aged 15–64 years at €7.9 billion. After adjusting for age, our estimated total cost was €16.3 billion; although about twice their estimate, the costs per case are comparable as our estimate was based on a total of 9.1 million cases.

Another source of uncertainty is the variation in healthcare systems between EU countries; we have taken account of this by averaging data and adjusting for relative average income levels. Taking the differences in levels of care and cost of each country’s healthcare system fully into account was beyond the scope of the present exercise. We used annual costs per patient together with prevalence rates of disease across Europe, both of which are inevitably associated with uncertainty. Only a minority of those with COPD have been diagnosed and, even though the undiagnosed population generally have very mild disease, undoubtedly some incur costs for healthcare and lost production not accounted for here.

In estimating the costs of lung cancer, the incidence approach is reasonable for small cell cancer with its average short survival, but is less appropriate for nonsmall cell lung cancer. The follow-up periods (up to 18 months and 30 months respectively) of the two lung cancer cost studies used may not have covered the full costs for these patients; thus we probably underestimated their healthcare costs. The indirect costs for lung cancer rely on a single cost estimate from Germany and assume that the incidence is representative of all EU countries.

For pneumonia/ALRI, our estimate serves as only a partial illustration of the economic burden, limited to costs of inpatient care. Healthcare and lost production costs were reported to cost about €10 billion annually in the previous edition of the White Book (2003), and in 2011 values this would be close to €12 billion. Although our estimate for healthcare is about a quarter of this, it is limited to inpatient costs. The cost of DALYs lost due to pneumonia, however, is included and is considerable.

Our estimate for OSAS relies on a single study from Denmark, and furthermore, the prevalence of OSAS for our calculations, 0.36%, was derived from the same study [32].

Extrapolation to the other 27 countries is inevitably associated with considerable uncertainty. Other studies generally report higher prevalence, albeit with considerable variation, related partly to varying definitions. A further important aspect of OSAS is its considerable socioeconomic impact beyond the usual direct and indirect costs, in particular related to road traffic accidents caused by the associated sleepiness. A review of obstructive sleep apnoea in the USA reported costs of some $16 billion (€17 billion at 2000 exchange rates) for road accidents in 2000.

The costs of cystic fibrosis are dependent on access to treatment and survival into adulthood, both of which may vary across Europe. We have no data on indirect costs, which are likely to be substantial due to premature death, disability and inability to work (including of parents), so once more our estimate is inevitably conservative.

The economic burden of bronchiectasis, pulmonary fibrosis, pulmonary vascular diseases, and occupational diseases such as asbestosis or silicosis could not be estimated, and would also add to the economic burden of respiratory disease, both from healthcare and work limitation such as lost production due to reduced efficiency at work, absence from work, early retirement and premature death; in 2004, new cases of occupational asthma in the United Kingdom alone cost £70–100 million (€115–165 billion at 2004 exchange rates).

Many of the causes of respiratory disease are behaviour related or otherwise potentially preventable, particularly those due to smoking tobacco or poor air quality. For example, it is estimated that 60% of COPD in the EU is attributable to smoking, along with 85% of lung cancer and 10% of other lower respiratory disease. This would suggest that the direct healthcare costs of respiratory disease attributable to smoking are approximately €27.4 billion, which is about half of the direct healthcare costs of respiratory disease estimated here. Consequently, with similar proportional savings in indirect costs and DALYs lost, a huge amount of money could be saved by reducing the preventable causes of respiratory disease in the EU.

The total cost of respiratory disease in EU, including the value of DALYs lost to respiratory disease, is estimated at a minimum of €380 billion. Even though this figure is a gross underestimate, as it excludes many respiratory diseases for which costs were not available, it represents a massive loss to the EU every year.

See the entire Economic Burden of Lung Disease Chapter