Clinical effectiveness
Extensive literature searches were conducted in an attempt to identify potentially relevant studies. We performed electronic searches of a range of bibliographic databases as well as screening of clinical trial registers. Conference proceedings were also searched to identify unpublished studies. The review process followed recommended methods to minimise any potential errors and biases. The quality of included studies was assessed in detail at outcome level and accounted for when interpreting the findings. Appropriate synthesis approaches were employed by taking into account the heterogeneity of study characteristics, and the meta-analyses adhered to a pre-defined analytic strategy.
In terms of limitations, only English-language studies were included; therefore, some potentially relevant non-English-language studies may have been missed. There was some evidence of inconsistency in the meta-analysis of mortality outcomes. A range of potential sources of heterogeneity were further explored. The observed heterogeneity may be explained by variations in trial quality, different risk profiles of
populations at baseline, and variations in the CT parameters used in included trials. In addition, there were wide variations in definitions of a positive scan on the lung nodule detection across trials.
Cost-effectiveness
This is a model developed independently by an experienced research group, free from potential conflicts of interest. It is also, we believe, the first economic evaluation of lung cancer screening to include a risk
prediction component with a variable threshold (but, for example, ten Haaf et al.115have used risk proxies
in the form of smoking histories).
The independent economic assessment evaluates the cost-effectiveness of a wide range of potential screening programmes, through the use of a natural history model (which allows for evaluation of hypothetical screening programmes that have not been evaluated in clinical trials). This natural history
model is based on high-quality evidence from the large NLST RCT71and UK national sources.1The
assessment also includes recent estimates of the costs of screening, and somewhat recent estimates of the cost of lung cancer. A clear description of the assumptions underpinning the assessment has been given. The economic evaluation seems to suggest with some robustness that screening is unlikely to be cost-effective at a threshold of £20,000 per QALY.
A number of assumptions were made in the construction of the economic model and some of these were explored in scenario or sensitivity analyses. No modelling of smoking behaviour was included, and incidental findings were not modelled. By using the DES framework, there has been no need to artificially restrict the model states or distributions for event times, or to consider a homogeneous cohort.
The model does not take the impact on mortality as an input, but produces it as an output resulting from the natural history model and the screening programme design. If additional mortality benefit (above what the model currently predicts) needs to be incorporated in an economic evaluation (i.e. if it is demonstrated in future data from trials), new assumptions and parameters will need to be introduced. This could be based on, for example, an acceleration factor applied to lead time. The current model predicts that the cost-effectiveness of screening is closely linked to the RR of lung cancer mortality (Figure 36), which suggests that, with a RR
of 0.935, single screening of individuals aged 60–75 years with ≥ 3% risk of lung cancer would become
cost-effective at £20,000 per QALY (although this is based on extrapolation and is therefore subject to significant uncertainty).
As a DES was used, greatly increasing the computational resource requirements for analyses, there is the risk that results are affected by Monte Carlo error because no inbuilt convergence checks were used, and certain analyses were conducted with a number of simulations known to be short of the apparent number required for stability.
DISCUSSION
NIHR Journals Library www.journalslibrary.nihr.ac.uk
Patient and public involvement
The diversity of perspectives and issues expressed by participants in discussions during our PPI consultations were audio-recorded and all meeting transcripts were analysed thematically to ensure an accurate and comprehensive record of our PPI consultation process for consideration in this and future related HTAs. A key strength of performing a thematic analysis of all workshop transcripts and constructing an explanatory model to reflect participant perspectives was that it ensured that these views were preserved throughout the research process and enabled efficient and accurate communication of PPI perspectives between PenTAG researchers, the majority of whom were not present at any of the PPI meetings. PenTAG researchers were consequently able to consider a variety of patient and public perspectives during the HTA process, particularly
issues and concerns relevant to‘at-risk’ asymptomatic smokers/former smokers recruited locally from deprived
areas. Patient and public views relating to the psychological impact of screening and HRQoL were referred to by PenTAG modellers while running scenario analyses and provided further assurance that the economic model analysis had face validity.
Conducting PPI meetings with a tailored workshop format designed specifically for this HTA ensured that views expressed in our PPI meetings were relevant to our HTA and the variety of perspectives and issues discussed reflect the complexity of real-world situations that could potentially affect uptake and
implementation of a potential UK lung cancer screening programme.
A key strength of our approach was that we were able to capture the perspectives of a range of patient and public members, with a particular focus on smokers/former smokers currently without symptoms of
lung cancer who may be considered‘high risk’ and, hence, a potential priority target group for a UK
national lung cancer screening programme. Views were particularly sought from smokers/former smokers recruited from local deprived areas. Our consultations were extended to include the views of non-smokers during a visit to a community centre/food bank in the most deprived ward in Exeter. At the end of each PPI workshop meeting, all participants were asked to comment on the findings of a recent qualitative study that investigated attitudes towards lung cancer screening in socioeconomically deprived and heavy-smoking
communities.170However, we acknowledge that many patients and members of the public have an interest
in this HTA and we were not able to include everyone. Owing to practical limitations (e.g. a notable lack of local established support groups for lung cancer patients/carers in our local Exeter area), we took the decision not to specifically recruit lung cancer patients/carers. However, people who lost close relatives to cancer (including lung cancer) and other smoking-related diseases were involved in our workshop meetings, as were people who had a family history of lung cancer.
100% 50% 0% y = – 464.02x + 433.71 – 45 – 40 – 35 – 30 – 25 – 20 – 15 – 10 – 5 0 0.94 0.95 0.96 0.97 0.98 0.99 1.00 1.01 1.02 1.03 INMB of S–60–75–3% vs. no screening (£)
Relative risk of lung cancer mortality in participants
FIGURE 36 Impact of RR of lung cancer mortality on cost-effectiveness. INMB at £20,000 per QALY.
DOI: 10.3310/hta22690 HEALTH TECHNOLOGY ASSESSMENT 2018 VOL. 22 NO. 69
© Queen’s Printer and Controller of HMSO 2018. This work was produced by Snowsill et al. under the terms of a commissioning contract issued by the Secretary of State for Health and Social Care. This issue may be freely reproduced for the purposes of private research and study and extracts (or indeed, the full report) may be included in professional journals provided that suitable acknowledgement is made and the reproduction is not associated with any form of advertising. Applications for commercial reproduction should be addressed to: NIHR Journals Library, National Institute for Health Research, Evaluation, Trials and Studies Coordinating Centre, Alpha House, University of Southampton Science Park, Southampton SO16 7NS, UK.