This study has shown that from a societal perspective the cost outcome ratio (COR) after implementing CBT for CFS in a MHC was dominant compared to before. From a healthcare perspective the COR after implementation was more costly but also more effective than before, and the 100% probability that the COR is acceptable was reached at the willingness to pay threshold of € 4.500 is positive. Given that CBT is the only effective treatment for CFS and has been scarcely available until now, this is relevant information in favor of nationwide implementation. Although some studies have already examined the cost effectiveness of behavioral treatments for chronic fatigue (CF) [7, 8] and for CFS , there has been no research into the cost effectiveness of such a treatment that also took into account the costs of designing the implementation interventions needed for implementing the treatment and the costs of actually implementing the treatment in a non-academic setting. Such a study implies a less homogenous patient population and less control over the content of performed treatment sessions than an academic setting can guarantee.
Concerning age and gender, the patient population was fully representative of the CFS population. Compared to other trials in the area of CFS, the baseline fatigue severity was a little lower and relatively many patients had a paid job [26, 27]. These differences could be explained by the fact that the treatment facility at the mental health care institution was more easily accessible. Patients may be recognized as CFS by their GP and referred to CBT in an earlier phase than patients referred (mostly by a medical specialist) to a specialized hospital setting.
As was also found in earlier cost effectiveness studies, [8, 9] an overall lower use of health care facilities was measured after CBT for CFS than before it. This may be explained by the fact that during treatment with CBT patients are instructed not to use other treatments or medication and by the fact that when starting treatment all patients were diagnosed as CFS. Looking for a diagnosis and a lack of affective treatment are the main reasons for CFS patients' high use of health care facilities . Concerning work productivity, fewer patients had a paid job after treatment than before, but the mean hours of paid work per week had increased after treatment. Given the short time horizon (8 months) the full influence of CBT for CFS on work productivity might be revealed to be larger and the impact on cost-effectiveness more pronounced.
In this study we used a conservative method, last observation carried forward, in cases of missing data. This imputation method might have influenced the results in a conservative, negative direction. However the proportion of missing data was in our opinion rather small (< 12%) thus the chance that significantly different results were obtained is small.
A serious limitation of this study is it's non-controlled before and after design, which implies that incremental cost effectiveness compared to a natural course control group, or compared to a guided support group controlling for any placebo effect, could not be analysed. However, the incremental cost effectiveness ratio (ICER) of CBT for CFS compared to usual care was recently reported by Severens et al. . The focus and contribution of the present study was primarily to investigate costs and consequences of implementing this evidence based treatment in a clinical practice setting. This is a relevant issue in bridging the gap between science and research, since proven (cost) effectiveness under laboratory conditions of RCTs does not guarantee the same in the practice field of health care. Both smaller treatment effects due to the less controlled situation and accompanying costs of including costs for implementing the treatment might change the cost-outcome ratio.
Another weak point in this study is the variable follow up time. Although the mean time period between intake and post treatment was 8.4 months, and analyses were done using this time horizon for all patients, the real time interval varied considerable. The problem hereby is that in fact we do not know what this implies for the results that were found.
A strong point though is the fact that, besides the usual included medical-, productivity-, and patient related costs also protocol driven- and implementation related costs were included , giving a more complete and more relevant view on the cost and outcomes of providing nationwide CBT for CFS.