Can CT screening for lung cancer in New Zealand be cost-effective?

Posted on by

Dr Richard Jaine, Dr Nisha Nair and Professor Tony Blakely

CT screening for lung cancer -1There is now strong evidence that screening for lung cancer with low-dose computed tomography (LDCT) scans is effective at reducing lung cancer mortality. So why aren’t countries rushing to introduce a screening programme? Because there is still doubt about its cost-effectiveness. In this blog, we discuss the uncertainties and suggest a way forward for New Zealand.

Lung cancer isn’t the most common cancer in New Zealand – it doesn’t make it into the “top three” cancers in men or women. But it is the most deadly cancer, accounting for almost 20% of all cancer deaths in New Zealand. Lung cancer carries such a high mortality because it is often picked up at a late stage, when the disease is advanced, has spread to other parts of the body, and is that much harder to treat. Consequently, there is a lot of interest in detecting lung cancer at its earlier stages.

Screening for lung cancer with three annual LDCT scans has been shown to reduce mortality from lung cancer (by ~20%) and overall mortality (by ~7%) compared to screening via chest x-ray (1). This US study (the National Lung Screening Trial or NLST) was conducted in more than 50,000 high risk people (defined as current or former smokers aged 55-74 years, with a 30 pack-year history and, if former smokers, quit within the last 15 years). Unsurprisingly, this finding has led to calls for funding for lung cancer screening.

CT screening for lung cancer -2In December 2013, the US Preventative Services Task Force recommended “annual screening for lung cancer with low-dose computed tomography (LDCT) in adults aged 55 to 80 years who have a 30 pack-year smoking history and currently smoke or have quit within the past 15 years” (2). In Europe, decisions have been more cautious. Decisions may be dependent on the results of the Dutch Belgian randomised lung cancer screening trial (NELSON), results of which are expected next year (3).

Closer to home, the Australians have also assessed the evidence surrounding lung cancer screening (4). They conclude that there are still several issues to resolve, including evaluating the costs and potential harms. And similarly, despite the positive efficacy findings from the NLST(1), the authors of the NLST study state:

“Before public policy recommendations are crafted, the cost-effectiveness of low-dose CT screening must be rigorously analysed.” And,

“The cost-effectiveness of low-dose CT screening must also be considered in the context of competing interventions, particularly smoking cessation.”

Gaps in the puzzle

There are certainly some knowledge gaps to fill before any kind of policy decision on a population-based LDCT screening programme for lung cancer is made in New Zealand. Cost-effectiveness is the only criteria we are focusing on in this blog, but there are others: alternative strategies that could also reduce the burden of lung cancer, the capacity of the New Zealand health system to deal with another screening programme, the potential impact on inequalities in lung cancer, among others (5).

So, what about the cost-effectiveness of LDCT screening?

Since the results of the NLST, there have been a number of studies published on the cost-effectiveness of LDCT screening for lung cancer. Perhaps unsurprisingly, there is marked variation and it is difficult to generalise and ‘guess’ at the potential cost-effectiveness in the New Zealand setting. In short, the table below highlights the (wide) range of findings from various recent studies.

Table 1 Summary of various studies examining the cost-effectiveness of LDCT screening for lung cancer.

Study and year

Country Target population Comparator used Cost per quality adjusted life year [1] (QALY) gained With smoking cessation programme
Black et al 2014(6) USA 55-74 year olds with 30 pack-year history No screening $NZ 119,000(95% uncertainty interval: $76,000 to $273,000)
Goffin et al 2015(7) Canada 55-74 year olds with 30 pack-year history No screening $NZ 61,000 $NZ 28,000
McMahon et al 2011(8) USA 50-74 year olds with 20 pack-year history No screening $NZ 187,000-251,000 $NZ 194,000-237,000
Shmeuli et al 2013(9) Israel 45 years or older, moderate-to-heavy smokers (median 37 pack-years) Usual care $NZ 2,148
Villanti et al 2013(10) USA 50-64 year olds with 30 pack-year history No screening $NZ 41,000 $NZ 24,000-34,000

Note: dollars have been converted to New Zealand dollars using purchasing power parities from OECD, by year of publication. Dollars rounded to nearest $1,000 except for Shmeuli et al. Cited costs per QALY gained are for LDCT only, not including (say) adjunct smoking cessation.

As can be seen in the table, there is clearly a great deal of variation regarding the cost-effectiveness. At the extremes, cost-effectiveness ranges from $NZ 2,148 to $NZ 251,000. Even within a single study, subgroup analysis had the cost per QALY ranging from $NZ 63,000 for current smokers to $NZ 903,000 for former smokers (6). This is by no means an exhaustive list of studies, but highlights the variation of the cost-effectiveness of lung cancer screening with LDCT.

What is causing the variation?

In general, the “bottom line” dollars per QALY in cost-effectiveness studies is far less important than in interrogating the assumptions and inputs that went into it. There are a range of reasons to explain the range of findings. Here are a few that may explain some of it:

  • Different interventions: Some of the studies have varied the intervention:
    • The study (9) with the lowest cost per QALY was based on a single baseline LDCT screening intervention, while the highest cost per QALY study(8) was based on annual LDCT screening.
    • A smoking cessation programme piggybacked onto the screening intervention can reduce the QALY gain by over 50% e.g. for the Goffin et al study (7) the cost per QALY fell from $NZ 61,000 to $NZ 28,000.
  • Different target populations: The population that is modelled will vary between the studies. While all use “high-risk” smokers over the age of 45, age bands and pack-years differ (see table above). We expect the most QALY gain will be in the 60-65 year age group, with reducing QALY gain for both the younger group (through lower risk of lung cancer) and the older groups (with less life expectancy to gain).
  • Different “effect size”: The assumed effectiveness of screening also varies between the studies. Studies use different assumptions regarding the proportion of early stage cancers identified in the screened population (generally assumed to be somewhere between 40% and 85%). As is expected, the higher this proportion, the more cost-effective screening will appear; the study with the lowest cost per QALY in the table above used 85% (9).
  • Proportion of overdiagnosis [2]: In screening, overdiagnosis includes abnormalities detected that are of questionable malignancy and cancers detected that would not have been diagnosed in the absence of screening. The proportion of overdiagnosis within the screened population is another parameter that varies greatly between the studies. Some studies assume up to 100% of the excess cases identified compared to radiography screening are due to overdiagnosis (6), while others either ignore overdiagnosis or assume it is 0% in their base-case. This is an area of genuine uncertainty, and will likely influence the cost-effectiveness of a programme. In the study that used different overdiagnosis estimates in the sensitivity analyses, the final cost per QALY reduced by more than 30% with lower overdiagnosis rates ($NZ 119,000 with 100% overdiagnosis to $NZ 81,000 with 50% overdiagnosis) (6).
  • Different assumptions around what impact a LDCT screening programme has on smoking cessation: There is genuine uncertainty regarding this as well. Studies have shown that smoking cessation rates could increase (11, 12), remain unchanged (11), or decrease if a screening programme is implemented (13). What is clear, however, is that change (or no change) in smoking cessation rates can affect the cost-effectiveness. Where modelled, an improvement in smoking cessation rates invariably led to a reduced cost per QALY; doubling the cessation rate can halve the cost per QALY (8).
  • Different baseline data: Baseline data (i.e. the incidence and mortality of lung cancer) used in studies will vary over time and place which will have a bearing on results. Importantly, the rate of lung cancer incidence varies between smoking groups in different studies – the higher the incidence rate, the more cost-effective the study is likely to be (as bigger gains are possible). No study, as yet, has used New Zealand specific lung cancer incidence and mortality data.

What should NZ do?

At first glance, this is not clear. But let’s break it down to steps we think should be undertaken.

Firstly, run a New Zealand-specific analysis. Use New Zealand-specific lung cancer rates, especially by ethnicity (lung cancer rates vary more by ethnicity in New Zealand than can be explained ‘simply’ by ex-, current- and never-smoker categories), and use New Zealand cost data.

Second, uptake matters. We know that breast and colorectal cancer screening uptake rates are lower for Māori. Given the high lung cancer incidence among Māori, it seems unlikely that a LDCT screening programme would generate less health gain per capita for Māori, but this needs to be evaluated.

We predict that modelling of LDCT screening will show it is cost-effective, especially for Māori smokers. However, there will still be genuine uncertainty and unintended consequences. For example, some might argue that LDCT screening will find other pathologies that save lives (or conversely that were irrelevant and just generate anxiety and costs). Acceptability and feasibility are all issues that – if initial modelling evaluations suggest LDCT screening is likely to be cost-effective – may suggest a pilot study may be warranted (a colorectal cancer screening ‘replay’). Debate will be needed as to whether NZ can ‘cope’ with another screening programme, or whether we should just redouble smoking cessation focus (essentially worrying about the lung cancer incidence two or more decades into the future, and not those about to present now).

However, to not assess the likely health gains, inequality impacts and cost of lung cancer screening may risk leaving the issue open to speculation and partially informed decision making.

[1] QALY or Quality-Adjusted Life Year: The remaining life expectancy, adjusted for quality of life. Think of one QALY as one year of life in perfect health.

[2] In screening, overdiagnosis includes abnormalities detected that are of questionable malignancy and cancers detected that would not have been diagnosed in the absence of screening

References

  1. Aberle D, Adams A, Berg C, Black W, Clapp J, Fagerstrom R, et al. Reduced Lung-Cancer Mortality with Low-Dose Computed Tomographic Screening. N Engl J Med. 2011;365(5):395-409.
  2. Moyer VA, U. S. Preventive Services Task Force. Screening for lung cancer: U.S. Preventive Services Task Force recommendation statement. Annals of Internal Medicine. 2014;160(5):330-8.
  3. Ru Zhao Y, Xie X, de Koning HJ, Mali WP, Vliegenthart R, Oudkerk M. NELSON lung cancer screening study. Cancer Imaging. 2011;11 Spec No A:S79-84.
  4. Department of Health and Aging. Lung Cancer Screening: Overview of the evidence and issues. Australia: Department of Health and Aging, 2013.
  5. National Health Committee. Screening to improve health in New Zealand: criteria to assess screening programmes. Wellington: National Health Committee, 2003.
  6. Black WC, Gareen IF, Soneji SS, Sicks JD, Keeler EB, Aberle DR, et al. Cost-Effectiveness of CT Screening in the National Lung Screening Trial. N Engl J Med. 2014;371(19):1793-802.
  7. Goffin JR, Flanagan WM, Miller AB, et al. Cost-effectiveness of lung cancer screening in canada. JAMA Oncology. 2015;1(6):807-13.
  8. McMahon PM, Kong CY, Bouzan C, Weinstein MC, Cipriano LE, Tramontano AC, et al. Cost-effectiveness of computed tomography screening for lung cancer in the United States. J Thorac Oncol. 2011;6(11):1841-8.
  9. Shmueli A, Fraifeld S, Peretz T, Gutfeld O, Gips M, Sosna J, et al. Cost-effectiveness of baseline low-dose computed tomography screening for lung cancer: the Israeli experience. Value in Health. 2013;16(6):922-31.
  10. Villanti AC, Jiang Y, Abrams DB, Pyenson BS. A Cost-Utility Analysis of Lung Cancer Screening and the Additional Benefits of Incorporating Smoking Cessation Interventions. PLoS ONE. 2013;8(8):e71379.
  11. Slatore CG, Baumann C, Pappas M, Humphrey LL. Smoking Behaviors among Patients Receiving Computed Tomography for Lung Cancer Screening. Systematic Review in Support of the U.S. Preventive Services Task Force. Ann Am Thorac Soc. 2014;11(4):619-27.
  12. Tammemagi MC, Berg CD, Riley TL, Cunningham CR, Taylor KL. Impact of lung cancer screening results on smoking cessation. J Natl Cancer Inst. 2014;106(6):dju084.
  13. van der Aalst CM, van den Bergh KAM, Willemsen MC, de Koning HJ, van Klaveren RJ. Lung cancer screening and smoking abstinence: 2 year follow-up data from the Dutch–Belgian randomised controlled lung cancer screening trial. Thorax. 2010;65(7):600-5.

This entry was posted in Uncategorized and tagged , , , by Kate Sloane. Bookmark the permalink.

5 thoughts on “Can CT screening for lung cancer in New Zealand be cost-effective?

  1. Thank you for initiating the discussion on lung cancer screening.

    There were a number of sessions on this topic in the recent Preventing Overdiagnosis conference in Washington and, and hopefully, some of the presentations will be available publicly at the website soon..
    http://www.preventingoverdiagnosis.net/

    Questions that were raised in the conference included the US preventive task force recommending lung cancer screening outside eligible age groups that were in the original trials as noted in the blog above. It was also questioned as to whether one should wait for more evidence as there are ongoing trials of lung cancer screening, again as noted in this blog.

    It was acknowledged that recommendations may be amended when new evidence becomes available.

    There were a number of concerns expressed in regard to harm related to lung cancer screening, including high rates of false positives, the number of people requiring following up and the high level of incidental findings. What is being observed in a real world observation study appears to be even higher than findings reported in the trials.

    There was a discussion about whether the CT scan should always ‘not include” the kidneys, or use a lower resolution scan (to reduce the rate of incidental findings).

    In regard to modelling the cost effectiveness of lung cancer screening in the NZ context, it would be useful to
    • consider the absolute benefit in terms of all cause mortality (as well as lung cancer mortality),
    • actively consider co-morbidities and the expected healthy life years for the cohort eligible for screening (as there are competing causes of death), and
    • the (ongoing) cost of incidental findings should also be included as part of the intervention pathway (this is often not included in most cost effectiveness models).

    As part of screening pathways one should also consider possible barriers to implementation such as radiology capacity and the workforce to provide appropriate follow up.

    Another important question to consider is what is the likely acceptability from the people who may be eligible for screening?

    Having a NNT as part of informed consent process is often helpful, in regard to benefit, as well as harms such as false positive, over treatment and over diagnosis. A lay language summary of such information would be very helpful.

    Finally, while we are discussing the merits and feasibility of lung cancer screening, it would be worth noting that this is happening passively in NZ in an informal way with the increase use of more advanced imaging, e.g. CT coronary angiograms are read by the cardiologists for the findings relevant to them and then the radiologists report the other areas. It would be helpful to review the cost effectiveness of such an approach, considering many people who have indications for CT coronary angiogram may not necessarily be heavy smokers, and what is the likelihood of false positive and incidental findings that may require surveillance or invasive investigation compared to the group formally eligible for lung cancer screening. Quantifying the benefits and harms of reporting findings of advanced imaging such as CT and MRI outside the indicated area of interest would be very interesting such as passive lung cancer screening or inadvertent screening for ‘incidentaloma’.

    • Kia ora Wing,

      Thanks for your very useful comments. It is great to have this dialogue as we are considering the parameters and inputs for the model. As mentioned in the blog there are clearly some uncertainties in trying to accurately model the cost-effectiveness of LDCT screening.

      I was also surprised that the US Preventive Task Force provided recommendations beyond the eligible age groups of the original NLST (55-74 years). Our baseline model would not go beyond the NLST age group (particularly given there is only limited data on screening for age groups beyond that).

      I agree with your suggestions for inputs into the cost-effectiveness model. I expect to be able to model all of those factors to some extent. Our model will also cover a range of levels of overdiagnosis, to cover the uncertainty that surrounds overdiagnosis.

      There are also plenty of other criteria to consider with screening programmes (you mention workforce capacity, acceptability, issues of informed consent). Obviously, the cost-effectiveness of the programme is only one part of this, so our work will be able to add to the debate in this – and we will also be able to add to debate on the other criteria. I would consider radiology capacity and diagnostic capacity are the major limiting factors at this stage (but not an insurmountable problem for the future). And as for informed consent – this needs to be improved across the board for current screening programmes in New Zealand, let alone the need to be thoughtfully considered for any proposed programmes.

      Thanks again for your comments.

      Cheers,
      Richard

    • Thanks, Wing, for raising some critical points.
      Completely agree that there are many things to consider before implementing such a screening program, and cost-effectiveness is just one of those considerations. You’ve mentioned several others: having a suitable test that is acceptable, ensuring the health system has the capacity to handle the entire CT screening pathway. The issue of false positives, over diagnosis, and incidental findings all speak to that trickiest of issues: the benefit of such a screening program versus the potential harms. The first part in dealing with that is to better define those harms (hopefully in ongoing work), and then dealing with how best to convey those harms to participants, as you’ve stated.
      While waiting for more good evidence to come in, there’s still a couple of things we can do from a modelling perspective: building the model in such a way that allows new evidence to be incorporated, and running what-if scenarios until such time as that evidence is available. Not perfect by a long shot, but a ‘best guess’!

  2. For those of us not familiar with the technique, what are the costs associated with low-dose computed tomography (LDCT) scanning and diagnosis?

    If it is a high up-front cost associated with capital intensive equipment, then these will be sensitive to discount rates, and the applicability of this equipment to other uses. (A CT machine able to be used for other programs may become more cost-effective). If the bulk of the program cost is due to; technicians, diagnostic staff, or administration and treatment, then the cost structure may be rather fixed. If there’s potential for automation and computer-assisted diagnostics then the cost may decline over time (but is likely to remain high in the immediate future).

    Those are very high cost-per-daly interventions, particularly compared to an increased tobacco tax or increasing the size of tobacco carton warnings, or targeted cessation programs. Or improvements in vehicle exhaust regulations.

    • Thanks for your comments George.
      You raise some good points about the costs of LDCT screening for lung cancer. At this early stage in our research, it is difficult to be precise about the costs of the screening pathway, although based on overseas research in this area and some local data, we expect the cost of a CT scan to be in the high hundreds of dollars. On top of that, there are initial physician consultation costs for a positive result (perhaps low hundreds), diagnostic procedures (costs would be quite variable depending on procedure including further CT scans, biopsies and bronchoscopies) and surgery costs (thousands of dollars), among others.
      While some of the other tobacco control interventions you mention (e.g. tobacco tax) have been shown to be cost effective in New Zealand, the cost-effectiveness of LDCT screening for lung cancer is not known for New Zealand – and by providing this information, we will be able to have more informed discussions about where we want to put the health dollar.
      Thanks again,
      Richard

Leave a Reply to Wing Cheuk Chan Cancel reply

Your email address will not be published. Required fields are marked *

* *