Discussion
The two most cost effective strategies, which also allowed histological confirmation of
the NSCLC, were identified from the literature. The study here suggests that PET is the more
cost effective than the CWU for both the CT negative and CT positive patients with ICERs of
$A 14,580 per QALY and $A 52,039 per QALY respectively. The CT positive ICER is high,
but there is no explicit threshold used in the Australian health care system for an acceptable
ICER, although, it is generally implicitly assumed to be around $A 50,000 dollars per QALY.
So whether this strategy would be used by a societal decision maker is questionable. However,
due to the large amount of variability in the strategy’s ICERs shown by the sensitivity analyses
this throws the strategy’s cost effectiveness seriously into doubt.
However, the sensitivity analyses show that for CT negative patients the results seem to
be extremely robust. The only sensitivity analysis that saw the ICER decrease noticeably was
when the rate of metastasis was decreased to 0.05 and then the ICER increased to $A 54,453
per QALY, otherwise it stayed within a range of $A 11,728 to $A 19,268 or it dominated the
CWU strategy.
This study largely confirms the results of the previous studies but in an Australian
context which has not been looked at before and it is thought that this is the first time that a
probabilistic sensitivity analysis has been carried out on PET looking at CT positive and CT
negative patients separately. Compared to the other studies there was more movement in the
results when the individual parameters were changed especially for the CT positive patients.
This no doubt reflects the substantial uncertainty over some of parameters in the model
especially concerning the underlying distribution of disease. This study has provided more
evidence suggesting that rate of metastasis is not the same for patients with different stages of
NSCLC. Further work on the quality of life estimates and the cost of radiotherapy and
chemotherapy in an Australian context would yield extra certainty in the reliability of these
results.
Despite the overwhelming international evidence and the now this study showing that
PET in CT negative patients is a cost effective and with greater uncertainty for CT positive
patients these results should still be interpreted cautiously due to many of the assumptions used
in the modelling. However, the recent introduction of a new generation of PET scanners and
the new combination CT-PET scanners that not only ensure better T staging, but improvements
in N and M staging. We should, therefore, start to see the cost effectiveness of PET improve
even further as scanning times are reduced and the accuracy improves .50
Acknowledgements
I would like to thank my supervisors, Richard Norman and Stephen Goodall with their help
and support and their comments on the drafts of this paper. Furthermore, I would like to thank
everyone at Centre for Health Economics Research and Evaluation (CHERE) for their
guidance and help in writing this paper.
50 Devaraj, A., G. J. Cook, et al. (2007). "PET/CT in non-small cell lung cancer staging-promises and problems."
Clinical Radiology 62(2): p.106
29