View Article PDF

In this viewpoint, we suggest that policymakers should prioritise health interventions by using evidence around health gain, impact on equity, health-system costs and cost-effectiveness. We take the example of the new Cancer Control Agency in New Zealand, Te Aho o Te Kahu, and argue that its decision-making can now be informed by many methodologically compatible epidemiological and health economic analyses. These analyses span primary prevention of cancer (eg, tobacco control, dietary and physical activity interventions and HPV vaccination), cancer screening, cancer treatment and palliative care. The largest health gain and cost-savings from the available modelling work for New Zealand are seen in nutrition and tobacco control interventions in particular. Many of these interventions have potentially greater per capita health gain for Māori than non-Māori and are also found to be cost saving for the health sector. In summary, appropriate prioritisation of interventions can potentially both maximise health benefits as well as making best use of government funding of the health system.

New Zealand now has the Cancer Control Agency Te Aho o Te Kahu (the ‘Agency’), which is committed to improving cancer control in the country. The Agency will need to determine its priorities for cancer control interventions and to consider such issues as the potential for health gain and the timing of those health gains, the potential benefit of reducing health inequities and getting the best value for money from health-system expenditure. We use the development of the Agency to argue in this viewpoint for improved use of modelling-based evidence to inform decision-making. There is a lot of evidence from health and medical science research about which interventions work (eg, randomised control trials), but seldom is the impact projected onto populations in the future. This is a glaring omission in much policy-making, which can be partially rectified through the quantification and comparison of metrics for specified interventions. We showcase examples of modelling studies of many interventions that could be taken into account in the Agency’s deliberations.

To illustrate available work for a New Zealand context, we extracted selected data from our online interactive league table (https://league-table.shinyapps.io/bode3/). Specifically, we extracted metrics for a range of potential interventions, from primary prevention of cancer to cancer screening, management and treatment and palliation. This is detailed in Table 1, including where cancer control was not the major cause of the estimated health gain, but where it contributes to at least some of the health gain. For example, tobacco control interventions typically generate more health gain from preventing chronic respiratory disease than cancer prevention, but cancer prevention is still an important component of the benefit. We present health gain and cost results for interventions over the remaining lifetime of the population modelled (generally the 2011 New Zealand population), discounted at 3% to account for greater societal value placed on health gains closer in time (unless otherwise stated). Quality-adjusted life years (QALYs) gained and 95% uncertainty intervals are shown in Figure 1 below.

Results of methodologically comparable New Zealand modelling studies of cancer control

Primary prevention via tobacco control

As per Table 1, the “sinking lid on tobacco supply” intervention is estimated to result in health gains of approximately 282,000 QALYs or 64 QALYs per 1,000 people, although only 31% of the health benefits were due to cancer prevention.1 This intervention may have a much higher QALY gain per capita for Māori than for non-Māori (at 156 versus 47 QALYs per 1,000 people gained, respectively).1

Tobacco tax increases of 10% annually to 2031 have also been estimated to result in large health gains (57,500 QALYs, or 13 QALYs per 1,000 people). All of the other tobacco control measures listed in Table 1 are cost-saving. The tobacco control measures listed are also estimated to result in higher per capita health gains for Māori than for non-Māori.

To be clear, it typically takes some time before the health benefit of such primary preventive interventions to peak. While some health gains occur within a few years, for certain interventions (eg, a tobacco tax) the actual (large) peak in health gains occurs in about 50 years (all other things held constant).

Primary prevention via nutrition and physical activity

Large health gains could be achieved through the adoption of climate-friendly, plant-based eating patterns. For example, the adoption of a waste-free vegan diet was estimated to result in 1.46 million QALYs gained (331 QALYs per 1,000 people).2 Specifically, the modelling included the prevention of 13 different diet-related cancers. This intervention was also estimated to be cost-saving and may have much higher per capita health gains for Māori than for non-Māori (508 versus 298 QALYs per 1,000 people, respectively).2 Although the adoption of a waste-free vegan diet may only be considered feasible for a minority of the population, even just shifting diets to meet New Zealand dietary guidelines was estimated to result in 1.02 million QALYs gained (232 QALYs per 1,000 people).2

A number of other dietary interventions may also result in substantial health gains and health system cost savings. For example, a fruit and vegetable subsidy in combination with a sugar tax was estimated to result in 894,000 QALYs gained (751 per 1,000 [undiscounted]).3 Other food taxes and subsidies (eg, just a sugar tax, just a fruit and vegetable subsidy) have also been assessed (Table 1) and, as was the case for the adoption of climate-friendly eating patterns, the modelling included the prevention of 13 different diet-related cancers.3

Most of the interventions related to nutrition and physical activity in Table 1 were estimated to be cost-saving, with the exception of weight-loss dietary counselling by nurses in primary care, mass media promotion of smartphone apps for weight-loss and mass media promotion of apps for physical activity. Most of the interventions presented in Table 1 may also result in higher per capita health gains for Māori than for non-Māori, with the exception of the mass media promotion of apps for weight-loss.4

Primary prevention: other domains

Of the other primary prevention measures presented in Table 1, an alcohol tax resulted in the highest health gains.5 This intervention was only modelled for transport injury prevention, but it would be expected to help prevent a range of alcohol-related cancers. It also had relatively low costs, although if a broader societal perspective was taken there would be NZ$240 million savings due to the reduction of social harms. HPV vaccination programmes could also provide health gains,6,7 with a mandatory school-based girls’ programme estimated to be more cost-effective than a programme for both sexes.6

Screening for cancer

Colorectal cancer screening has been estimated to result in health gains of approximately 101,800 QALYs and was considered cost-effective.8 Helicobacter pylori (H pylori) faecal antigen and serology screening at the national level have been estimated to result in health gains of 15,300 and 14,200 QALYs from stomach cancer prevention, respectively.9 Although H pylori screening programmes were estimated to achieve much less health gain than many other prevention programmes, the health benefits typically occur sooner and were cost-effective for Māori and Pacific people. Both H pylori faecal antigen and serology screening would potentially result in much higher per capita health gains for Māori than for non-Māori. However, serology screening was estimated to be more cost-effective, with a lower incremental cost-effectiveness ratio (ICER) than faecal antigen screening.

Low-dose CT screening for lung cancer was found to possibly be cost-effectiveness for Māori in our analysis (using a threshold of $45,000 per QALY gained, a rule of thumb of GDP per capita per QALY gained).10 An updated analysis using the same model11 reported screening to be more cost-effective and also cost-effective for non-Māori. This model included NELSON randomised controlled trial (RCT) findings,12 assumed an equal Māori and non-Māori participation in screening and assumed that low-dose CT screening achieves the same stage distribution as in RCTs (rather than a proportionate shift in stage distribution from New Zealand’s relatively poor current stage distribution). It is moot whether a low-dose CT screening programme in New Zealand can achieve equal coverage and the same stage distribution as in RCTs; the most sensible way forward may be a pilot study in New Zealand to test these assumptions.

Management and treatment of cancer

With regards to the interventions assessed for managing cancer treatment presented in Table 1, cancer care coordinators for colorectal cancer patients would likely result in the largest health gains (84 QALYs) and was the most cost-effective intervention. However, it is critical to note that these evaluations differ from those detailed above in that they are ‘just’ for people diagnosed with cancer in one year, and the health gains typically occur within several years rather than decades. If the intervention had been modelled every year for the next 20 years, and if there had been no changes in other interventions (albeit this is unlikely) and no change in population size, then the health gains would be roughly 15 times greater (less than 20 due to discounting at 3%). In addition, the intervention is likely to have higher per capita health gains for Māori than for non-Māori. The health equity evidence for the other interventions for managing cancer treatment was less clear.

Palliation

An economic evaluation of single-fraction (SFX) versus multiple-fraction (MFX) palliative radiotherapy for painful bone metastases in breast, lung and prostate cancer found that, although QALY gains were similar for SFX and MFX, the per patient costs were less for SFX.13

Table 1: Impact of various cancer control interventions according to methodologically compatible BODE3 epidemiological and health economic modelling (all over the remaining life course of the modelled population, at 3% discount rate, NZ$ 2016; the published works in journals and league table provide uncertainty intervals; ordering of tabulated interventions is in terms of declining health gain within each subsection).

* Health gain presented in the published paper as HALYs. HALYs are health-adjusted life-years and can be considered equivalent to QALYs in this case.

Figure 1: Health gains (QALYs) for selected cancer control interventions with 3% discounting (bars show 95% uncertainty intervals).

Discussion

A particular strength of this ‘league table’ with these modelled cancer-related interventions is that all the modelling used similar epidemiological/costing data and generally the same modelling structure and methods (a proportional multistate lifetable, as stated in Table 1), albeit with some fairly minor differences for displaying some results (eg, costing in NZ$ 2011 vs 2016 currency rates). Another strength is that there are interventions that span the full range of cancer domains from primary prevention to screening, treatment and palliation.

Furthermore, modelling is typically far more feasible and lower cost than running RCTs in New Zealand (each with an associated health economic analysis) for each plausible intervention. Indeed, RCTs are not readily feasible for studying certain policy interventions (eg, tobacco tax increases or changes to food industry practices).

In terms of the size of health benefits, primary prevention interventions tend to have much larger health gains (in QALYs) relative to screening, management and treatment and palliative interventions. However, prevention’s gains are usually many years later, and the population coverage of an intervention (eg, population-wide or at-risk groups), intervention duration and discount rate can contribute significantly to the size of the health benefits. Primary prevention interventions are also more likely to be cost-saving for the health system, and so can potentially ‘liberate’ government funds for other uses. In particular, there is a strong case for the prioritisation of enhancing tobacco control interventions (which are also compatible with the Smokefree 2025 Goal held by a succession of New Zealand governments). Furthermore, all of the tobacco control interventions included in Table 1 are estimated to result in greater per capita gain for Māori than non-Māori.

There is also a potentially strong case for investment in diet-related interventions. Improving nutrition is estimated to typically produce large health gains and to save costs (see Table 1). In addition, almost all of the dietary interventions included had higher per capita health gains for Māori than non-Māori. Nutrition interventions may also have many co-benefits that are not accounted for in the modelling work (eg, reduced damage to the environment from greenhouse gases and water pollution from more plant-based diets). Some of these nutrition interventions can have widespread public support depending on how they are framed. For example, a sugary drinks tax can potentially be highly favoured if the tax revenue is used to fund child health28 (eg, healthy school lunches).

However, there may be substantial political feasibility constraints with some tobacco, nutrition and alcohol control interventions. In comparison, for some other cancer prevention interventions there would probably be higher public support (eg, raising the levels of HPV vaccination at school up to the higher levels observed for Australia and the UK). Improved targeting of cancer treatment by cancer subtype (eg, as per our study on Herceptin for breast cancer)26 is also likely to be fairly non-controversial as it is an appropriate use of ‘personalised medicine’ and makes best use of available resources.29,30

Possible next steps

Given this wealth of methodologically compatible data, we argue that agencies such as Te Aho o Te Kahu should give consideration to routinely using it in their decision-making processes. They will also need to bring other factors into the prioritisation process, as does PHARMAC, which has a prioritisation framework31 and which makes use of epidemiological and health economic modelling. In addition to considering health gain, health inequities, cost and cost-effectiveness, other key factors in prioritisation that health agencies need to consider include:

  • Political commitment and, in particular, how health interventions rank compared to the major policy items competing for legislative time in Parliament. Political engagement can also be impacted by the ‘rescue prerogative’, whereby the benefits of identifiably sick individuals (eg, for access to new cancer medicines) are sometimes given political priority over saving statistical lives. Concern for this prerogative can be seen as a democratic government responding to the demands of its citizens, but it can also be seen as a driver of cost-ineffective use of government funds and a driver of inequities.
  • Intervention feasibility (eg, especially where there is opposition from vested commercial interests as per the tobacco industry).
  • Upfront costs, which are especially relevant in the constrained COVID-19-related fiscal environment. However, policymakers should ideally take a long-term view so that the benefits, which may peak decades in the future (eg, from tobacco control and improved nutrition), are accounted for.
  • Co-benefits that are outside of a health-system perspective. These include impact of interventions on preventing income loss, preventing other societal harms (eg, the impact of alcohol on crime) and preventing greenhouse gases (eg, via dietary and transport-related interventions).
  • Effect sizes from RCTs that are used (wherever possible) are obviously important for the robustness of model outputs. Where evaluations are based on effect sizes not generated from RCTs, it is important to note that modelling is typically far more feasible and lower cost than running RCTs in New Zealand; if the evaluation clearly shows the intervention is not cost-effective or highly cost-effective for plausible assumptions about key effect sizes, then no RCT is needed. However, if the result is equivocal, the imperative for RCTs (if possible and feasible to conduct) for critical input parameters increases.

Agencies should consider commissioning additional analyses and looking at methodologically compatible modelling results for Australian-based interventions (ie, also in our online league table that includes compatible New Zealand and Australian studies: https://league-table.shinyapps.io/bode3/).

Limitations of using models to inform decision-making

Although we favour the use of quantification and modelling studies to inform decision-making, we note important issues around uncertainty. Specifically, there is uncertainty around all modelling results, and we have attempted to quantify this in the 95% uncertainty intervals for all the key results, including the QALYs (see Figure 1), as well as the costs and ICERs (see the published papers referenced in Table 1). These papers also typically have sensitivity and scenario analyses that allow the reader to consider how the results may vary with different assumptions and parameter values. Nevertheless, there are additional areas of uncertainty that are harder to quantify, including model structure uncertainty (eg, in the ideal world different research teams would independently build models of the same interventions). There are also assumptions around business-as-usual trends, and such assumptions can be made less valid by sudden health and economic shocks (eg, as per the COVID-19 pandemic). And finally, for some hypothetical interventions (eg, a sinking lid on tobacco supply that has never been fully realised in any jurisdiction in the world) there are multiple unknowns that may go beyond the uncertainty estimates we have made. Despite these issues, epidemiological and health economic modelling provides quantitative insights that should be the starting point to inform decision-making in our view.

Summary

Abstract

In this viewpoint, we suggest that policymakers should prioritise health interventions by using evidence around health gain, impact on equity, health-system costs and cost-effectiveness. We take the example of the new Cancer Control Agency in New Zealand, Te Aho o Te Kahu, and argue that its decision-making can now be informed by many methodologically compatible epidemiological and health economic analyses. These analyses span primary prevention of cancer (eg, tobacco control, dietary and physical activity interventions and HPV vaccination), cancer screening, cancer treatment and palliative care. The largest health gain and cost-savings from the available modelling work for New Zealand are seen in nutrition and tobacco control interventions in particular. Many of these interventions have potentially greater per capita health gain for Māori than non-Māori and are also found to be cost saving for the health sector. In summary, appropriate prioritisation of interventions can potentially both maximise health benefits as well as making best use of government funding of the health system.

Aim

Method

Results

Conclusion

Author Information

Prof Nick Wilson: BODE3 Programme, University of Otago Wellington, New Zealand. Dr Leah Grout: BODE3 Programme, University of Otago Wellington, New Zealand. Dr Jennifer Summers: BODE3 Programme, University of Otago Wellington, New Zealand. Dr Amanda C Jones: BODE3 Programme, University of Otago Wellington, New Zealand. Dr Anja Mizdrak: BODE3 Programme, University of Otago Wellington, New Zealand. Dr Nhung Nghiem: BODE3 Programme, University of Otago Wellington, New Zealand. Dr Cristina Cleghorn: BODE3 Programme, University of Otago Wellington, New Zealand. Prof Tony Blakely: BODE3 Programme, University of Otago Wellington, New Zealand; Population Interventions, Centre for Epidemiology and Biostatistics, Melbourne School of Population and Global Health, University of Melbourne, Melbourne, Australia.

Acknowledgements

This BODE3 modelling work has been supported by the Health Research Council of New Zealand (Grants 10/248 and 16/443) and by the Ministry of Business Innovation and Employment (MBIE) (Grant UOOX1406). The authors thank all former and current colleagues in BODE3 for contributing to the body of work referred to in this article.

Correspondence

Dr Leah Grout, University of Otago

Correspondence Email

leah.grout@otago.ac.nz

Competing Interests

Nil.

1. van der Deen FS, Wilson N, Cleghorn CL, Kvizhinadze G, Cobiac LJ, Nghiem N, Blakely T. Impact of five tobacco endgame strategies on future smoking prevalence, population health and health system costs: two modelling studies to inform the tobacco endgame. Tob Control 2018;27:278-86.

2. Drew J, Cleghorn C, Macmillan A, Mizdrak A. Healthy and Climate-Friendly Eating Patterns in the New Zealand Context. Environ Health Perspect 2020;128:17007.

3. Blakely T, Cleghorn C, Mizdrak A, Waterlander W, Nghiem N, Swinburn B, Wilson N, Ni Mhurchu C. The effect of food taxes and subsidies on population health and health costs: a modelling study. Lancet Public Health 2020;5:e404-e13.

4. Cleghorn C, Wilson N, Nair N, Kvizhinadze G, Nghiem N, McLeod M, Blakely T. Health Benefits and Cost-Effectiveness From Promoting Smartphone Apps for Weight Loss: Multistate Life Table Modeling. JMIR Mhealth Uhealth 2019;7:e11118.

5. Cobiac LJ, Mizdrak A, Wilson N. Cost-effectiveness of raising alcohol excise taxes to reduce the injury burden of road traffic crashes. Inj Prev 2019;25:421-27.

6. Pearson AL, Kvizhinadze G, Wilson N, Smith M, Canfell K, Blakely T. Is expanding HPV vaccination programs to include school-aged boys likely to be value-for-money: a cost-utility analysis in a country with an existing school-girl program. BMC Infect Dis 2014;14:351.

7. Blakely T, Kvizhinadze G, Karvonen T, Pearson AL, Smith M, Wilson N. Cost-effectiveness and equity impacts of three HPV vaccination programmes for school-aged girls in New Zealand. Vaccine 2014;32:2645-56.

8. McLeod M, Kvizhinadze G, Boyd M, Barendregt J, Sarfati D, Wilson N, Blakely T. Colorectal cancer screening: How health gains and cost-effectiveness vary by ethnic group, the impact on health inequalities, and the optimal age range to screen. Cancer Epidemiol Biomarkers Prev 2017;26:1391-400.

9. Teng AM, Kvizhinadze G, Nair N, McLeod M, Wilson N, Blakely T. A screening program to test and treat for Helicobacter pylori infection: Cost-utility analysis by age, sex and ethnicity. BMC Infect Dis 2017;17:156.

10. Jaine R, Kvizhinadze G, Nair N, Blakely T. Cost-effectiveness of a low-dose computed tomography screening programme for lung cancer in New Zealand. Lung Cancer 2020;144:99-106.

11. McLeod M, Sandiford P, Kvizhinadze G, Bartholomew K, Crengle S. Impact of low-dose CT screening for lung cancer on ethnic health inequities in New Zealand: a cost-effectiveness analysis. BMJ Open 2020;10:e037145.

12. Horeweg N, Scholten ET, de Jong PA, van der Aalst CM, Weenink C, Lammers JW, Nackaerts K, Vliegenthart R, ten Haaf K, Yousaf-Khan UA, Heuvelmans MA, Thunnissen E, Oudkerk M, Mali W, de Koning HJ. Detection of lung cancer through low-dose CT screening (NELSON): a prespecified analysis of screening test performance and interval cancers. Lancet Oncol 2014;15:1342-50.

13. Collinson L, Kvizhinadze G, Nair N, McLeod M, Blakely T. Economic evaluation of single-fraction versus multiple-fraction palliative radiotherapy for painful bone metastases in breast, lung and prostate cancer. J Med Imaging Radiat Oncol 2016;60:650-60.

14. Blakely T, Cobiac LJ, Cleghorn CL, Pearson AL, van der Deen FS, Kvizhinadze G, Nghiem N, McLeod M, Wilson N. Health, health inequality, and cost impacts of annual increases in tobacco tax: Multistate life table modeling in New Zealand. PLoS Med 2015;12:e1001856.

15. Blakely T, Cobiac LJ, Cleghorn CL, Pearson AL, van der Deen FS, Kvizhinadze G, Nghiem N, McLeod M, Wilson N. Correction: Health, Health Inequality, and Cost Impacts of Annual Increases in Tobacco Tax: Multistate Life Table Modeling in New Zealand. PLoS Med 2016;13:e1002211.

16. Nghiem N, Leung W, Cleghorn C, Blakely T, Wilson N. Mass media promotion of a smartphone smoking cessation app: modelled health and cost-saving impacts. BMC Public Health 2019;19:283.

17. Nghiem N, Cleghorn CL, Leung W, Nair N, van der Deen FS, Blakely T, Wilson N. A national quitline service and its promotion in the mass media: modelling the health gain, health equity and cost-utility. Tob Control 2018;27:434–41.

18. Nghiem N, Blakely T, Cobiac LJ, Cleghorn CL, Wilson N. The health gains and cost savings of dietary salt reduction interventions, with equity and age distributional aspects. BMC Public Health 2016;16:423.

19. Cleghorn C, Blakely T, Jones A, Kvizhinadze G, Mizdrak A, Nghiem N, Ni Mhurchu C, Wilson N. Feasible diet intervention options to improve health and save costs for the New Zealand population. Burden of Disease Epidemiology, Equity and Cost-Effectiveness Programme. Wellington: Department of Public Health, University of Otago, Wellington, 2019. https://www.otago.ac.nz/wellington/departments/publichealth/research/bode3/publications/#reports. 2019.

20. Mizdrak A, Blakely T, Cleghorn CL, Cobiac LJ. Potential of active transport to improve health, reduce healthcare costs, and reduce greenhouse gas emissions: A modelling study. PLoS One 2019;14:e0219316.

21. Cleghorn C, Blakely T, Mhurchu CN, Wilson N, Neal B, Eyles H. Estimating the health benefits and cost-savings of a cap on the size of single serve sugar-sweetened beverages. Prev Med 2019;120:150-56.

22. Cleghorn CL, Wilson N, Nair N, Kvizhinadze G, Nghiem N, McLeod M, Blakely T. Health benefits and costs of weight-loss dietary counselling by nurses in primary care: a cost-effectiveness analysis. Public Health Nutr 2020;23:83-93.

23. Mizdrak A, Telfer K, Direito A, Cobiac LJ, Blakely T, Cleghorn CL, Wilson N. Health Gain, Cost Impacts, and Cost-Effectiveness of a Mass Media Campaign to Promote Smartphone Apps for Physical Activity: Modeling Study. JMIR Mhealth Uhealth 2020;8:e18014.

24. Blakely T, Collinson L, Kvizhinadze G, Nair N, Foster R, Dennett E, Sarfati D. Cancer care coordinators in stage III colon cancer: a cost-utility analysis. BMC Health Serv Res 2015;15:306.

25. Webber-Foster R, Kvizhinadze G, Rivalland G, Blakely T. Cost-effectiveness analysis of docetaxel versus weekly paclitaxel in adjuvant treatment of regional breast cancer in New Zealand. Pharmacoeconomics 2014;32:707-24.

26. Leung W, Kvizhinadze G, Nair N, Blakely T. Adjuvant Trastuzumab in HER2-Positive Early Breast Cancer by Age and Hormone Receptor Status: A Cost-Utility Analysis. PLoS Med 2016;13:e1002067.

27. Nair N, Kvizhinadze G, Blakely T. Cancer Care Coordinators to Improve Tamoxifen Persistence in Breast Cancer: How Heterogeneity in Baseline Prognosis Impacts on Cost-Effectiveness. Value Health 2016;19:936-44.

28. Sundborn G. Policy brief: a sugary drink tax for New Zealand and 10,000-strong petition snubbed by Minister of Health and National Government. N Z Med J 2017;130:114-16.

29. Jackson SE, Chester JD. Personalised cancer medicine. Int J Cancer 2015;137:262-66.

30. Haslem DS, Van Norman SB, Fulde G, Knighton AJ, Belnap T, Butler AM, Rhagunath S, Newman D, Gilbert H, Tudor BP, Lin K, Stone GR, Loughmiller DL, Mishra PJ, Srivastava R, Ford JM, Nadauld LD. A Retrospective Analysis of Precision Medicine Outcomes in Patients With Advanced Cancer Reveals Improved Progression-Free Survival Without Increased Health Care Costs. J Oncol Pract 2017;13:e108-e19.

31. PHARMAC. Making funding decisions (Fact Sheet #4). Wellington: PHARMAC. https://www.pharmac.govt.nz/assets/factsheet-04-making-funding-decisions.pdf.

Contact diana@nzma.org.nz
for the PDF of this article

View Article PDF

In this viewpoint, we suggest that policymakers should prioritise health interventions by using evidence around health gain, impact on equity, health-system costs and cost-effectiveness. We take the example of the new Cancer Control Agency in New Zealand, Te Aho o Te Kahu, and argue that its decision-making can now be informed by many methodologically compatible epidemiological and health economic analyses. These analyses span primary prevention of cancer (eg, tobacco control, dietary and physical activity interventions and HPV vaccination), cancer screening, cancer treatment and palliative care. The largest health gain and cost-savings from the available modelling work for New Zealand are seen in nutrition and tobacco control interventions in particular. Many of these interventions have potentially greater per capita health gain for Māori than non-Māori and are also found to be cost saving for the health sector. In summary, appropriate prioritisation of interventions can potentially both maximise health benefits as well as making best use of government funding of the health system.

New Zealand now has the Cancer Control Agency Te Aho o Te Kahu (the ‘Agency’), which is committed to improving cancer control in the country. The Agency will need to determine its priorities for cancer control interventions and to consider such issues as the potential for health gain and the timing of those health gains, the potential benefit of reducing health inequities and getting the best value for money from health-system expenditure. We use the development of the Agency to argue in this viewpoint for improved use of modelling-based evidence to inform decision-making. There is a lot of evidence from health and medical science research about which interventions work (eg, randomised control trials), but seldom is the impact projected onto populations in the future. This is a glaring omission in much policy-making, which can be partially rectified through the quantification and comparison of metrics for specified interventions. We showcase examples of modelling studies of many interventions that could be taken into account in the Agency’s deliberations.

To illustrate available work for a New Zealand context, we extracted selected data from our online interactive league table (https://league-table.shinyapps.io/bode3/). Specifically, we extracted metrics for a range of potential interventions, from primary prevention of cancer to cancer screening, management and treatment and palliation. This is detailed in Table 1, including where cancer control was not the major cause of the estimated health gain, but where it contributes to at least some of the health gain. For example, tobacco control interventions typically generate more health gain from preventing chronic respiratory disease than cancer prevention, but cancer prevention is still an important component of the benefit. We present health gain and cost results for interventions over the remaining lifetime of the population modelled (generally the 2011 New Zealand population), discounted at 3% to account for greater societal value placed on health gains closer in time (unless otherwise stated). Quality-adjusted life years (QALYs) gained and 95% uncertainty intervals are shown in Figure 1 below.

Results of methodologically comparable New Zealand modelling studies of cancer control

Primary prevention via tobacco control

As per Table 1, the “sinking lid on tobacco supply” intervention is estimated to result in health gains of approximately 282,000 QALYs or 64 QALYs per 1,000 people, although only 31% of the health benefits were due to cancer prevention.1 This intervention may have a much higher QALY gain per capita for Māori than for non-Māori (at 156 versus 47 QALYs per 1,000 people gained, respectively).1

Tobacco tax increases of 10% annually to 2031 have also been estimated to result in large health gains (57,500 QALYs, or 13 QALYs per 1,000 people). All of the other tobacco control measures listed in Table 1 are cost-saving. The tobacco control measures listed are also estimated to result in higher per capita health gains for Māori than for non-Māori.

To be clear, it typically takes some time before the health benefit of such primary preventive interventions to peak. While some health gains occur within a few years, for certain interventions (eg, a tobacco tax) the actual (large) peak in health gains occurs in about 50 years (all other things held constant).

Primary prevention via nutrition and physical activity

Large health gains could be achieved through the adoption of climate-friendly, plant-based eating patterns. For example, the adoption of a waste-free vegan diet was estimated to result in 1.46 million QALYs gained (331 QALYs per 1,000 people).2 Specifically, the modelling included the prevention of 13 different diet-related cancers. This intervention was also estimated to be cost-saving and may have much higher per capita health gains for Māori than for non-Māori (508 versus 298 QALYs per 1,000 people, respectively).2 Although the adoption of a waste-free vegan diet may only be considered feasible for a minority of the population, even just shifting diets to meet New Zealand dietary guidelines was estimated to result in 1.02 million QALYs gained (232 QALYs per 1,000 people).2

A number of other dietary interventions may also result in substantial health gains and health system cost savings. For example, a fruit and vegetable subsidy in combination with a sugar tax was estimated to result in 894,000 QALYs gained (751 per 1,000 [undiscounted]).3 Other food taxes and subsidies (eg, just a sugar tax, just a fruit and vegetable subsidy) have also been assessed (Table 1) and, as was the case for the adoption of climate-friendly eating patterns, the modelling included the prevention of 13 different diet-related cancers.3

Most of the interventions related to nutrition and physical activity in Table 1 were estimated to be cost-saving, with the exception of weight-loss dietary counselling by nurses in primary care, mass media promotion of smartphone apps for weight-loss and mass media promotion of apps for physical activity. Most of the interventions presented in Table 1 may also result in higher per capita health gains for Māori than for non-Māori, with the exception of the mass media promotion of apps for weight-loss.4

Primary prevention: other domains

Of the other primary prevention measures presented in Table 1, an alcohol tax resulted in the highest health gains.5 This intervention was only modelled for transport injury prevention, but it would be expected to help prevent a range of alcohol-related cancers. It also had relatively low costs, although if a broader societal perspective was taken there would be NZ$240 million savings due to the reduction of social harms. HPV vaccination programmes could also provide health gains,6,7 with a mandatory school-based girls’ programme estimated to be more cost-effective than a programme for both sexes.6

Screening for cancer

Colorectal cancer screening has been estimated to result in health gains of approximately 101,800 QALYs and was considered cost-effective.8 Helicobacter pylori (H pylori) faecal antigen and serology screening at the national level have been estimated to result in health gains of 15,300 and 14,200 QALYs from stomach cancer prevention, respectively.9 Although H pylori screening programmes were estimated to achieve much less health gain than many other prevention programmes, the health benefits typically occur sooner and were cost-effective for Māori and Pacific people. Both H pylori faecal antigen and serology screening would potentially result in much higher per capita health gains for Māori than for non-Māori. However, serology screening was estimated to be more cost-effective, with a lower incremental cost-effectiveness ratio (ICER) than faecal antigen screening.

Low-dose CT screening for lung cancer was found to possibly be cost-effectiveness for Māori in our analysis (using a threshold of $45,000 per QALY gained, a rule of thumb of GDP per capita per QALY gained).10 An updated analysis using the same model11 reported screening to be more cost-effective and also cost-effective for non-Māori. This model included NELSON randomised controlled trial (RCT) findings,12 assumed an equal Māori and non-Māori participation in screening and assumed that low-dose CT screening achieves the same stage distribution as in RCTs (rather than a proportionate shift in stage distribution from New Zealand’s relatively poor current stage distribution). It is moot whether a low-dose CT screening programme in New Zealand can achieve equal coverage and the same stage distribution as in RCTs; the most sensible way forward may be a pilot study in New Zealand to test these assumptions.

Management and treatment of cancer

With regards to the interventions assessed for managing cancer treatment presented in Table 1, cancer care coordinators for colorectal cancer patients would likely result in the largest health gains (84 QALYs) and was the most cost-effective intervention. However, it is critical to note that these evaluations differ from those detailed above in that they are ‘just’ for people diagnosed with cancer in one year, and the health gains typically occur within several years rather than decades. If the intervention had been modelled every year for the next 20 years, and if there had been no changes in other interventions (albeit this is unlikely) and no change in population size, then the health gains would be roughly 15 times greater (less than 20 due to discounting at 3%). In addition, the intervention is likely to have higher per capita health gains for Māori than for non-Māori. The health equity evidence for the other interventions for managing cancer treatment was less clear.

Palliation

An economic evaluation of single-fraction (SFX) versus multiple-fraction (MFX) palliative radiotherapy for painful bone metastases in breast, lung and prostate cancer found that, although QALY gains were similar for SFX and MFX, the per patient costs were less for SFX.13

Table 1: Impact of various cancer control interventions according to methodologically compatible BODE3 epidemiological and health economic modelling (all over the remaining life course of the modelled population, at 3% discount rate, NZ$ 2016; the published works in journals and league table provide uncertainty intervals; ordering of tabulated interventions is in terms of declining health gain within each subsection).

* Health gain presented in the published paper as HALYs. HALYs are health-adjusted life-years and can be considered equivalent to QALYs in this case.

Figure 1: Health gains (QALYs) for selected cancer control interventions with 3% discounting (bars show 95% uncertainty intervals).

Discussion

A particular strength of this ‘league table’ with these modelled cancer-related interventions is that all the modelling used similar epidemiological/costing data and generally the same modelling structure and methods (a proportional multistate lifetable, as stated in Table 1), albeit with some fairly minor differences for displaying some results (eg, costing in NZ$ 2011 vs 2016 currency rates). Another strength is that there are interventions that span the full range of cancer domains from primary prevention to screening, treatment and palliation.

Furthermore, modelling is typically far more feasible and lower cost than running RCTs in New Zealand (each with an associated health economic analysis) for each plausible intervention. Indeed, RCTs are not readily feasible for studying certain policy interventions (eg, tobacco tax increases or changes to food industry practices).

In terms of the size of health benefits, primary prevention interventions tend to have much larger health gains (in QALYs) relative to screening, management and treatment and palliative interventions. However, prevention’s gains are usually many years later, and the population coverage of an intervention (eg, population-wide or at-risk groups), intervention duration and discount rate can contribute significantly to the size of the health benefits. Primary prevention interventions are also more likely to be cost-saving for the health system, and so can potentially ‘liberate’ government funds for other uses. In particular, there is a strong case for the prioritisation of enhancing tobacco control interventions (which are also compatible with the Smokefree 2025 Goal held by a succession of New Zealand governments). Furthermore, all of the tobacco control interventions included in Table 1 are estimated to result in greater per capita gain for Māori than non-Māori.

There is also a potentially strong case for investment in diet-related interventions. Improving nutrition is estimated to typically produce large health gains and to save costs (see Table 1). In addition, almost all of the dietary interventions included had higher per capita health gains for Māori than non-Māori. Nutrition interventions may also have many co-benefits that are not accounted for in the modelling work (eg, reduced damage to the environment from greenhouse gases and water pollution from more plant-based diets). Some of these nutrition interventions can have widespread public support depending on how they are framed. For example, a sugary drinks tax can potentially be highly favoured if the tax revenue is used to fund child health28 (eg, healthy school lunches).

However, there may be substantial political feasibility constraints with some tobacco, nutrition and alcohol control interventions. In comparison, for some other cancer prevention interventions there would probably be higher public support (eg, raising the levels of HPV vaccination at school up to the higher levels observed for Australia and the UK). Improved targeting of cancer treatment by cancer subtype (eg, as per our study on Herceptin for breast cancer)26 is also likely to be fairly non-controversial as it is an appropriate use of ‘personalised medicine’ and makes best use of available resources.29,30

Possible next steps

Given this wealth of methodologically compatible data, we argue that agencies such as Te Aho o Te Kahu should give consideration to routinely using it in their decision-making processes. They will also need to bring other factors into the prioritisation process, as does PHARMAC, which has a prioritisation framework31 and which makes use of epidemiological and health economic modelling. In addition to considering health gain, health inequities, cost and cost-effectiveness, other key factors in prioritisation that health agencies need to consider include:

  • Political commitment and, in particular, how health interventions rank compared to the major policy items competing for legislative time in Parliament. Political engagement can also be impacted by the ‘rescue prerogative’, whereby the benefits of identifiably sick individuals (eg, for access to new cancer medicines) are sometimes given political priority over saving statistical lives. Concern for this prerogative can be seen as a democratic government responding to the demands of its citizens, but it can also be seen as a driver of cost-ineffective use of government funds and a driver of inequities.
  • Intervention feasibility (eg, especially where there is opposition from vested commercial interests as per the tobacco industry).
  • Upfront costs, which are especially relevant in the constrained COVID-19-related fiscal environment. However, policymakers should ideally take a long-term view so that the benefits, which may peak decades in the future (eg, from tobacco control and improved nutrition), are accounted for.
  • Co-benefits that are outside of a health-system perspective. These include impact of interventions on preventing income loss, preventing other societal harms (eg, the impact of alcohol on crime) and preventing greenhouse gases (eg, via dietary and transport-related interventions).
  • Effect sizes from RCTs that are used (wherever possible) are obviously important for the robustness of model outputs. Where evaluations are based on effect sizes not generated from RCTs, it is important to note that modelling is typically far more feasible and lower cost than running RCTs in New Zealand; if the evaluation clearly shows the intervention is not cost-effective or highly cost-effective for plausible assumptions about key effect sizes, then no RCT is needed. However, if the result is equivocal, the imperative for RCTs (if possible and feasible to conduct) for critical input parameters increases.

Agencies should consider commissioning additional analyses and looking at methodologically compatible modelling results for Australian-based interventions (ie, also in our online league table that includes compatible New Zealand and Australian studies: https://league-table.shinyapps.io/bode3/).

Limitations of using models to inform decision-making

Although we favour the use of quantification and modelling studies to inform decision-making, we note important issues around uncertainty. Specifically, there is uncertainty around all modelling results, and we have attempted to quantify this in the 95% uncertainty intervals for all the key results, including the QALYs (see Figure 1), as well as the costs and ICERs (see the published papers referenced in Table 1). These papers also typically have sensitivity and scenario analyses that allow the reader to consider how the results may vary with different assumptions and parameter values. Nevertheless, there are additional areas of uncertainty that are harder to quantify, including model structure uncertainty (eg, in the ideal world different research teams would independently build models of the same interventions). There are also assumptions around business-as-usual trends, and such assumptions can be made less valid by sudden health and economic shocks (eg, as per the COVID-19 pandemic). And finally, for some hypothetical interventions (eg, a sinking lid on tobacco supply that has never been fully realised in any jurisdiction in the world) there are multiple unknowns that may go beyond the uncertainty estimates we have made. Despite these issues, epidemiological and health economic modelling provides quantitative insights that should be the starting point to inform decision-making in our view.

Summary

Abstract

In this viewpoint, we suggest that policymakers should prioritise health interventions by using evidence around health gain, impact on equity, health-system costs and cost-effectiveness. We take the example of the new Cancer Control Agency in New Zealand, Te Aho o Te Kahu, and argue that its decision-making can now be informed by many methodologically compatible epidemiological and health economic analyses. These analyses span primary prevention of cancer (eg, tobacco control, dietary and physical activity interventions and HPV vaccination), cancer screening, cancer treatment and palliative care. The largest health gain and cost-savings from the available modelling work for New Zealand are seen in nutrition and tobacco control interventions in particular. Many of these interventions have potentially greater per capita health gain for Māori than non-Māori and are also found to be cost saving for the health sector. In summary, appropriate prioritisation of interventions can potentially both maximise health benefits as well as making best use of government funding of the health system.

Aim

Method

Results

Conclusion

Author Information

Prof Nick Wilson: BODE3 Programme, University of Otago Wellington, New Zealand. Dr Leah Grout: BODE3 Programme, University of Otago Wellington, New Zealand. Dr Jennifer Summers: BODE3 Programme, University of Otago Wellington, New Zealand. Dr Amanda C Jones: BODE3 Programme, University of Otago Wellington, New Zealand. Dr Anja Mizdrak: BODE3 Programme, University of Otago Wellington, New Zealand. Dr Nhung Nghiem: BODE3 Programme, University of Otago Wellington, New Zealand. Dr Cristina Cleghorn: BODE3 Programme, University of Otago Wellington, New Zealand. Prof Tony Blakely: BODE3 Programme, University of Otago Wellington, New Zealand; Population Interventions, Centre for Epidemiology and Biostatistics, Melbourne School of Population and Global Health, University of Melbourne, Melbourne, Australia.

Acknowledgements

This BODE3 modelling work has been supported by the Health Research Council of New Zealand (Grants 10/248 and 16/443) and by the Ministry of Business Innovation and Employment (MBIE) (Grant UOOX1406). The authors thank all former and current colleagues in BODE3 for contributing to the body of work referred to in this article.

Correspondence

Dr Leah Grout, University of Otago

Correspondence Email

leah.grout@otago.ac.nz

Competing Interests

Nil.

1. van der Deen FS, Wilson N, Cleghorn CL, Kvizhinadze G, Cobiac LJ, Nghiem N, Blakely T. Impact of five tobacco endgame strategies on future smoking prevalence, population health and health system costs: two modelling studies to inform the tobacco endgame. Tob Control 2018;27:278-86.

2. Drew J, Cleghorn C, Macmillan A, Mizdrak A. Healthy and Climate-Friendly Eating Patterns in the New Zealand Context. Environ Health Perspect 2020;128:17007.

3. Blakely T, Cleghorn C, Mizdrak A, Waterlander W, Nghiem N, Swinburn B, Wilson N, Ni Mhurchu C. The effect of food taxes and subsidies on population health and health costs: a modelling study. Lancet Public Health 2020;5:e404-e13.

4. Cleghorn C, Wilson N, Nair N, Kvizhinadze G, Nghiem N, McLeod M, Blakely T. Health Benefits and Cost-Effectiveness From Promoting Smartphone Apps for Weight Loss: Multistate Life Table Modeling. JMIR Mhealth Uhealth 2019;7:e11118.

5. Cobiac LJ, Mizdrak A, Wilson N. Cost-effectiveness of raising alcohol excise taxes to reduce the injury burden of road traffic crashes. Inj Prev 2019;25:421-27.

6. Pearson AL, Kvizhinadze G, Wilson N, Smith M, Canfell K, Blakely T. Is expanding HPV vaccination programs to include school-aged boys likely to be value-for-money: a cost-utility analysis in a country with an existing school-girl program. BMC Infect Dis 2014;14:351.

7. Blakely T, Kvizhinadze G, Karvonen T, Pearson AL, Smith M, Wilson N. Cost-effectiveness and equity impacts of three HPV vaccination programmes for school-aged girls in New Zealand. Vaccine 2014;32:2645-56.

8. McLeod M, Kvizhinadze G, Boyd M, Barendregt J, Sarfati D, Wilson N, Blakely T. Colorectal cancer screening: How health gains and cost-effectiveness vary by ethnic group, the impact on health inequalities, and the optimal age range to screen. Cancer Epidemiol Biomarkers Prev 2017;26:1391-400.

9. Teng AM, Kvizhinadze G, Nair N, McLeod M, Wilson N, Blakely T. A screening program to test and treat for Helicobacter pylori infection: Cost-utility analysis by age, sex and ethnicity. BMC Infect Dis 2017;17:156.

10. Jaine R, Kvizhinadze G, Nair N, Blakely T. Cost-effectiveness of a low-dose computed tomography screening programme for lung cancer in New Zealand. Lung Cancer 2020;144:99-106.

11. McLeod M, Sandiford P, Kvizhinadze G, Bartholomew K, Crengle S. Impact of low-dose CT screening for lung cancer on ethnic health inequities in New Zealand: a cost-effectiveness analysis. BMJ Open 2020;10:e037145.

12. Horeweg N, Scholten ET, de Jong PA, van der Aalst CM, Weenink C, Lammers JW, Nackaerts K, Vliegenthart R, ten Haaf K, Yousaf-Khan UA, Heuvelmans MA, Thunnissen E, Oudkerk M, Mali W, de Koning HJ. Detection of lung cancer through low-dose CT screening (NELSON): a prespecified analysis of screening test performance and interval cancers. Lancet Oncol 2014;15:1342-50.

13. Collinson L, Kvizhinadze G, Nair N, McLeod M, Blakely T. Economic evaluation of single-fraction versus multiple-fraction palliative radiotherapy for painful bone metastases in breast, lung and prostate cancer. J Med Imaging Radiat Oncol 2016;60:650-60.

14. Blakely T, Cobiac LJ, Cleghorn CL, Pearson AL, van der Deen FS, Kvizhinadze G, Nghiem N, McLeod M, Wilson N. Health, health inequality, and cost impacts of annual increases in tobacco tax: Multistate life table modeling in New Zealand. PLoS Med 2015;12:e1001856.

15. Blakely T, Cobiac LJ, Cleghorn CL, Pearson AL, van der Deen FS, Kvizhinadze G, Nghiem N, McLeod M, Wilson N. Correction: Health, Health Inequality, and Cost Impacts of Annual Increases in Tobacco Tax: Multistate Life Table Modeling in New Zealand. PLoS Med 2016;13:e1002211.

16. Nghiem N, Leung W, Cleghorn C, Blakely T, Wilson N. Mass media promotion of a smartphone smoking cessation app: modelled health and cost-saving impacts. BMC Public Health 2019;19:283.

17. Nghiem N, Cleghorn CL, Leung W, Nair N, van der Deen FS, Blakely T, Wilson N. A national quitline service and its promotion in the mass media: modelling the health gain, health equity and cost-utility. Tob Control 2018;27:434–41.

18. Nghiem N, Blakely T, Cobiac LJ, Cleghorn CL, Wilson N. The health gains and cost savings of dietary salt reduction interventions, with equity and age distributional aspects. BMC Public Health 2016;16:423.

19. Cleghorn C, Blakely T, Jones A, Kvizhinadze G, Mizdrak A, Nghiem N, Ni Mhurchu C, Wilson N. Feasible diet intervention options to improve health and save costs for the New Zealand population. Burden of Disease Epidemiology, Equity and Cost-Effectiveness Programme. Wellington: Department of Public Health, University of Otago, Wellington, 2019. https://www.otago.ac.nz/wellington/departments/publichealth/research/bode3/publications/#reports. 2019.

20. Mizdrak A, Blakely T, Cleghorn CL, Cobiac LJ. Potential of active transport to improve health, reduce healthcare costs, and reduce greenhouse gas emissions: A modelling study. PLoS One 2019;14:e0219316.

21. Cleghorn C, Blakely T, Mhurchu CN, Wilson N, Neal B, Eyles H. Estimating the health benefits and cost-savings of a cap on the size of single serve sugar-sweetened beverages. Prev Med 2019;120:150-56.

22. Cleghorn CL, Wilson N, Nair N, Kvizhinadze G, Nghiem N, McLeod M, Blakely T. Health benefits and costs of weight-loss dietary counselling by nurses in primary care: a cost-effectiveness analysis. Public Health Nutr 2020;23:83-93.

23. Mizdrak A, Telfer K, Direito A, Cobiac LJ, Blakely T, Cleghorn CL, Wilson N. Health Gain, Cost Impacts, and Cost-Effectiveness of a Mass Media Campaign to Promote Smartphone Apps for Physical Activity: Modeling Study. JMIR Mhealth Uhealth 2020;8:e18014.

24. Blakely T, Collinson L, Kvizhinadze G, Nair N, Foster R, Dennett E, Sarfati D. Cancer care coordinators in stage III colon cancer: a cost-utility analysis. BMC Health Serv Res 2015;15:306.

25. Webber-Foster R, Kvizhinadze G, Rivalland G, Blakely T. Cost-effectiveness analysis of docetaxel versus weekly paclitaxel in adjuvant treatment of regional breast cancer in New Zealand. Pharmacoeconomics 2014;32:707-24.

26. Leung W, Kvizhinadze G, Nair N, Blakely T. Adjuvant Trastuzumab in HER2-Positive Early Breast Cancer by Age and Hormone Receptor Status: A Cost-Utility Analysis. PLoS Med 2016;13:e1002067.

27. Nair N, Kvizhinadze G, Blakely T. Cancer Care Coordinators to Improve Tamoxifen Persistence in Breast Cancer: How Heterogeneity in Baseline Prognosis Impacts on Cost-Effectiveness. Value Health 2016;19:936-44.

28. Sundborn G. Policy brief: a sugary drink tax for New Zealand and 10,000-strong petition snubbed by Minister of Health and National Government. N Z Med J 2017;130:114-16.

29. Jackson SE, Chester JD. Personalised cancer medicine. Int J Cancer 2015;137:262-66.

30. Haslem DS, Van Norman SB, Fulde G, Knighton AJ, Belnap T, Butler AM, Rhagunath S, Newman D, Gilbert H, Tudor BP, Lin K, Stone GR, Loughmiller DL, Mishra PJ, Srivastava R, Ford JM, Nadauld LD. A Retrospective Analysis of Precision Medicine Outcomes in Patients With Advanced Cancer Reveals Improved Progression-Free Survival Without Increased Health Care Costs. J Oncol Pract 2017;13:e108-e19.

31. PHARMAC. Making funding decisions (Fact Sheet #4). Wellington: PHARMAC. https://www.pharmac.govt.nz/assets/factsheet-04-making-funding-decisions.pdf.

Contact diana@nzma.org.nz
for the PDF of this article

View Article PDF

In this viewpoint, we suggest that policymakers should prioritise health interventions by using evidence around health gain, impact on equity, health-system costs and cost-effectiveness. We take the example of the new Cancer Control Agency in New Zealand, Te Aho o Te Kahu, and argue that its decision-making can now be informed by many methodologically compatible epidemiological and health economic analyses. These analyses span primary prevention of cancer (eg, tobacco control, dietary and physical activity interventions and HPV vaccination), cancer screening, cancer treatment and palliative care. The largest health gain and cost-savings from the available modelling work for New Zealand are seen in nutrition and tobacco control interventions in particular. Many of these interventions have potentially greater per capita health gain for Māori than non-Māori and are also found to be cost saving for the health sector. In summary, appropriate prioritisation of interventions can potentially both maximise health benefits as well as making best use of government funding of the health system.

New Zealand now has the Cancer Control Agency Te Aho o Te Kahu (the ‘Agency’), which is committed to improving cancer control in the country. The Agency will need to determine its priorities for cancer control interventions and to consider such issues as the potential for health gain and the timing of those health gains, the potential benefit of reducing health inequities and getting the best value for money from health-system expenditure. We use the development of the Agency to argue in this viewpoint for improved use of modelling-based evidence to inform decision-making. There is a lot of evidence from health and medical science research about which interventions work (eg, randomised control trials), but seldom is the impact projected onto populations in the future. This is a glaring omission in much policy-making, which can be partially rectified through the quantification and comparison of metrics for specified interventions. We showcase examples of modelling studies of many interventions that could be taken into account in the Agency’s deliberations.

To illustrate available work for a New Zealand context, we extracted selected data from our online interactive league table (https://league-table.shinyapps.io/bode3/). Specifically, we extracted metrics for a range of potential interventions, from primary prevention of cancer to cancer screening, management and treatment and palliation. This is detailed in Table 1, including where cancer control was not the major cause of the estimated health gain, but where it contributes to at least some of the health gain. For example, tobacco control interventions typically generate more health gain from preventing chronic respiratory disease than cancer prevention, but cancer prevention is still an important component of the benefit. We present health gain and cost results for interventions over the remaining lifetime of the population modelled (generally the 2011 New Zealand population), discounted at 3% to account for greater societal value placed on health gains closer in time (unless otherwise stated). Quality-adjusted life years (QALYs) gained and 95% uncertainty intervals are shown in Figure 1 below.

Results of methodologically comparable New Zealand modelling studies of cancer control

Primary prevention via tobacco control

As per Table 1, the “sinking lid on tobacco supply” intervention is estimated to result in health gains of approximately 282,000 QALYs or 64 QALYs per 1,000 people, although only 31% of the health benefits were due to cancer prevention.1 This intervention may have a much higher QALY gain per capita for Māori than for non-Māori (at 156 versus 47 QALYs per 1,000 people gained, respectively).1

Tobacco tax increases of 10% annually to 2031 have also been estimated to result in large health gains (57,500 QALYs, or 13 QALYs per 1,000 people). All of the other tobacco control measures listed in Table 1 are cost-saving. The tobacco control measures listed are also estimated to result in higher per capita health gains for Māori than for non-Māori.

To be clear, it typically takes some time before the health benefit of such primary preventive interventions to peak. While some health gains occur within a few years, for certain interventions (eg, a tobacco tax) the actual (large) peak in health gains occurs in about 50 years (all other things held constant).

Primary prevention via nutrition and physical activity

Large health gains could be achieved through the adoption of climate-friendly, plant-based eating patterns. For example, the adoption of a waste-free vegan diet was estimated to result in 1.46 million QALYs gained (331 QALYs per 1,000 people).2 Specifically, the modelling included the prevention of 13 different diet-related cancers. This intervention was also estimated to be cost-saving and may have much higher per capita health gains for Māori than for non-Māori (508 versus 298 QALYs per 1,000 people, respectively).2 Although the adoption of a waste-free vegan diet may only be considered feasible for a minority of the population, even just shifting diets to meet New Zealand dietary guidelines was estimated to result in 1.02 million QALYs gained (232 QALYs per 1,000 people).2

A number of other dietary interventions may also result in substantial health gains and health system cost savings. For example, a fruit and vegetable subsidy in combination with a sugar tax was estimated to result in 894,000 QALYs gained (751 per 1,000 [undiscounted]).3 Other food taxes and subsidies (eg, just a sugar tax, just a fruit and vegetable subsidy) have also been assessed (Table 1) and, as was the case for the adoption of climate-friendly eating patterns, the modelling included the prevention of 13 different diet-related cancers.3

Most of the interventions related to nutrition and physical activity in Table 1 were estimated to be cost-saving, with the exception of weight-loss dietary counselling by nurses in primary care, mass media promotion of smartphone apps for weight-loss and mass media promotion of apps for physical activity. Most of the interventions presented in Table 1 may also result in higher per capita health gains for Māori than for non-Māori, with the exception of the mass media promotion of apps for weight-loss.4

Primary prevention: other domains

Of the other primary prevention measures presented in Table 1, an alcohol tax resulted in the highest health gains.5 This intervention was only modelled for transport injury prevention, but it would be expected to help prevent a range of alcohol-related cancers. It also had relatively low costs, although if a broader societal perspective was taken there would be NZ$240 million savings due to the reduction of social harms. HPV vaccination programmes could also provide health gains,6,7 with a mandatory school-based girls’ programme estimated to be more cost-effective than a programme for both sexes.6

Screening for cancer

Colorectal cancer screening has been estimated to result in health gains of approximately 101,800 QALYs and was considered cost-effective.8 Helicobacter pylori (H pylori) faecal antigen and serology screening at the national level have been estimated to result in health gains of 15,300 and 14,200 QALYs from stomach cancer prevention, respectively.9 Although H pylori screening programmes were estimated to achieve much less health gain than many other prevention programmes, the health benefits typically occur sooner and were cost-effective for Māori and Pacific people. Both H pylori faecal antigen and serology screening would potentially result in much higher per capita health gains for Māori than for non-Māori. However, serology screening was estimated to be more cost-effective, with a lower incremental cost-effectiveness ratio (ICER) than faecal antigen screening.

Low-dose CT screening for lung cancer was found to possibly be cost-effectiveness for Māori in our analysis (using a threshold of $45,000 per QALY gained, a rule of thumb of GDP per capita per QALY gained).10 An updated analysis using the same model11 reported screening to be more cost-effective and also cost-effective for non-Māori. This model included NELSON randomised controlled trial (RCT) findings,12 assumed an equal Māori and non-Māori participation in screening and assumed that low-dose CT screening achieves the same stage distribution as in RCTs (rather than a proportionate shift in stage distribution from New Zealand’s relatively poor current stage distribution). It is moot whether a low-dose CT screening programme in New Zealand can achieve equal coverage and the same stage distribution as in RCTs; the most sensible way forward may be a pilot study in New Zealand to test these assumptions.

Management and treatment of cancer

With regards to the interventions assessed for managing cancer treatment presented in Table 1, cancer care coordinators for colorectal cancer patients would likely result in the largest health gains (84 QALYs) and was the most cost-effective intervention. However, it is critical to note that these evaluations differ from those detailed above in that they are ‘just’ for people diagnosed with cancer in one year, and the health gains typically occur within several years rather than decades. If the intervention had been modelled every year for the next 20 years, and if there had been no changes in other interventions (albeit this is unlikely) and no change in population size, then the health gains would be roughly 15 times greater (less than 20 due to discounting at 3%). In addition, the intervention is likely to have higher per capita health gains for Māori than for non-Māori. The health equity evidence for the other interventions for managing cancer treatment was less clear.

Palliation

An economic evaluation of single-fraction (SFX) versus multiple-fraction (MFX) palliative radiotherapy for painful bone metastases in breast, lung and prostate cancer found that, although QALY gains were similar for SFX and MFX, the per patient costs were less for SFX.13

Table 1: Impact of various cancer control interventions according to methodologically compatible BODE3 epidemiological and health economic modelling (all over the remaining life course of the modelled population, at 3% discount rate, NZ$ 2016; the published works in journals and league table provide uncertainty intervals; ordering of tabulated interventions is in terms of declining health gain within each subsection).

* Health gain presented in the published paper as HALYs. HALYs are health-adjusted life-years and can be considered equivalent to QALYs in this case.

Figure 1: Health gains (QALYs) for selected cancer control interventions with 3% discounting (bars show 95% uncertainty intervals).

Discussion

A particular strength of this ‘league table’ with these modelled cancer-related interventions is that all the modelling used similar epidemiological/costing data and generally the same modelling structure and methods (a proportional multistate lifetable, as stated in Table 1), albeit with some fairly minor differences for displaying some results (eg, costing in NZ$ 2011 vs 2016 currency rates). Another strength is that there are interventions that span the full range of cancer domains from primary prevention to screening, treatment and palliation.

Furthermore, modelling is typically far more feasible and lower cost than running RCTs in New Zealand (each with an associated health economic analysis) for each plausible intervention. Indeed, RCTs are not readily feasible for studying certain policy interventions (eg, tobacco tax increases or changes to food industry practices).

In terms of the size of health benefits, primary prevention interventions tend to have much larger health gains (in QALYs) relative to screening, management and treatment and palliative interventions. However, prevention’s gains are usually many years later, and the population coverage of an intervention (eg, population-wide or at-risk groups), intervention duration and discount rate can contribute significantly to the size of the health benefits. Primary prevention interventions are also more likely to be cost-saving for the health system, and so can potentially ‘liberate’ government funds for other uses. In particular, there is a strong case for the prioritisation of enhancing tobacco control interventions (which are also compatible with the Smokefree 2025 Goal held by a succession of New Zealand governments). Furthermore, all of the tobacco control interventions included in Table 1 are estimated to result in greater per capita gain for Māori than non-Māori.

There is also a potentially strong case for investment in diet-related interventions. Improving nutrition is estimated to typically produce large health gains and to save costs (see Table 1). In addition, almost all of the dietary interventions included had higher per capita health gains for Māori than non-Māori. Nutrition interventions may also have many co-benefits that are not accounted for in the modelling work (eg, reduced damage to the environment from greenhouse gases and water pollution from more plant-based diets). Some of these nutrition interventions can have widespread public support depending on how they are framed. For example, a sugary drinks tax can potentially be highly favoured if the tax revenue is used to fund child health28 (eg, healthy school lunches).

However, there may be substantial political feasibility constraints with some tobacco, nutrition and alcohol control interventions. In comparison, for some other cancer prevention interventions there would probably be higher public support (eg, raising the levels of HPV vaccination at school up to the higher levels observed for Australia and the UK). Improved targeting of cancer treatment by cancer subtype (eg, as per our study on Herceptin for breast cancer)26 is also likely to be fairly non-controversial as it is an appropriate use of ‘personalised medicine’ and makes best use of available resources.29,30

Possible next steps

Given this wealth of methodologically compatible data, we argue that agencies such as Te Aho o Te Kahu should give consideration to routinely using it in their decision-making processes. They will also need to bring other factors into the prioritisation process, as does PHARMAC, which has a prioritisation framework31 and which makes use of epidemiological and health economic modelling. In addition to considering health gain, health inequities, cost and cost-effectiveness, other key factors in prioritisation that health agencies need to consider include:

  • Political commitment and, in particular, how health interventions rank compared to the major policy items competing for legislative time in Parliament. Political engagement can also be impacted by the ‘rescue prerogative’, whereby the benefits of identifiably sick individuals (eg, for access to new cancer medicines) are sometimes given political priority over saving statistical lives. Concern for this prerogative can be seen as a democratic government responding to the demands of its citizens, but it can also be seen as a driver of cost-ineffective use of government funds and a driver of inequities.
  • Intervention feasibility (eg, especially where there is opposition from vested commercial interests as per the tobacco industry).
  • Upfront costs, which are especially relevant in the constrained COVID-19-related fiscal environment. However, policymakers should ideally take a long-term view so that the benefits, which may peak decades in the future (eg, from tobacco control and improved nutrition), are accounted for.
  • Co-benefits that are outside of a health-system perspective. These include impact of interventions on preventing income loss, preventing other societal harms (eg, the impact of alcohol on crime) and preventing greenhouse gases (eg, via dietary and transport-related interventions).
  • Effect sizes from RCTs that are used (wherever possible) are obviously important for the robustness of model outputs. Where evaluations are based on effect sizes not generated from RCTs, it is important to note that modelling is typically far more feasible and lower cost than running RCTs in New Zealand; if the evaluation clearly shows the intervention is not cost-effective or highly cost-effective for plausible assumptions about key effect sizes, then no RCT is needed. However, if the result is equivocal, the imperative for RCTs (if possible and feasible to conduct) for critical input parameters increases.

Agencies should consider commissioning additional analyses and looking at methodologically compatible modelling results for Australian-based interventions (ie, also in our online league table that includes compatible New Zealand and Australian studies: https://league-table.shinyapps.io/bode3/).

Limitations of using models to inform decision-making

Although we favour the use of quantification and modelling studies to inform decision-making, we note important issues around uncertainty. Specifically, there is uncertainty around all modelling results, and we have attempted to quantify this in the 95% uncertainty intervals for all the key results, including the QALYs (see Figure 1), as well as the costs and ICERs (see the published papers referenced in Table 1). These papers also typically have sensitivity and scenario analyses that allow the reader to consider how the results may vary with different assumptions and parameter values. Nevertheless, there are additional areas of uncertainty that are harder to quantify, including model structure uncertainty (eg, in the ideal world different research teams would independently build models of the same interventions). There are also assumptions around business-as-usual trends, and such assumptions can be made less valid by sudden health and economic shocks (eg, as per the COVID-19 pandemic). And finally, for some hypothetical interventions (eg, a sinking lid on tobacco supply that has never been fully realised in any jurisdiction in the world) there are multiple unknowns that may go beyond the uncertainty estimates we have made. Despite these issues, epidemiological and health economic modelling provides quantitative insights that should be the starting point to inform decision-making in our view.

Summary

Abstract

In this viewpoint, we suggest that policymakers should prioritise health interventions by using evidence around health gain, impact on equity, health-system costs and cost-effectiveness. We take the example of the new Cancer Control Agency in New Zealand, Te Aho o Te Kahu, and argue that its decision-making can now be informed by many methodologically compatible epidemiological and health economic analyses. These analyses span primary prevention of cancer (eg, tobacco control, dietary and physical activity interventions and HPV vaccination), cancer screening, cancer treatment and palliative care. The largest health gain and cost-savings from the available modelling work for New Zealand are seen in nutrition and tobacco control interventions in particular. Many of these interventions have potentially greater per capita health gain for Māori than non-Māori and are also found to be cost saving for the health sector. In summary, appropriate prioritisation of interventions can potentially both maximise health benefits as well as making best use of government funding of the health system.

Aim

Method

Results

Conclusion

Author Information

Prof Nick Wilson: BODE3 Programme, University of Otago Wellington, New Zealand. Dr Leah Grout: BODE3 Programme, University of Otago Wellington, New Zealand. Dr Jennifer Summers: BODE3 Programme, University of Otago Wellington, New Zealand. Dr Amanda C Jones: BODE3 Programme, University of Otago Wellington, New Zealand. Dr Anja Mizdrak: BODE3 Programme, University of Otago Wellington, New Zealand. Dr Nhung Nghiem: BODE3 Programme, University of Otago Wellington, New Zealand. Dr Cristina Cleghorn: BODE3 Programme, University of Otago Wellington, New Zealand. Prof Tony Blakely: BODE3 Programme, University of Otago Wellington, New Zealand; Population Interventions, Centre for Epidemiology and Biostatistics, Melbourne School of Population and Global Health, University of Melbourne, Melbourne, Australia.

Acknowledgements

This BODE3 modelling work has been supported by the Health Research Council of New Zealand (Grants 10/248 and 16/443) and by the Ministry of Business Innovation and Employment (MBIE) (Grant UOOX1406). The authors thank all former and current colleagues in BODE3 for contributing to the body of work referred to in this article.

Correspondence

Dr Leah Grout, University of Otago

Correspondence Email

leah.grout@otago.ac.nz

Competing Interests

Nil.

1. van der Deen FS, Wilson N, Cleghorn CL, Kvizhinadze G, Cobiac LJ, Nghiem N, Blakely T. Impact of five tobacco endgame strategies on future smoking prevalence, population health and health system costs: two modelling studies to inform the tobacco endgame. Tob Control 2018;27:278-86.

2. Drew J, Cleghorn C, Macmillan A, Mizdrak A. Healthy and Climate-Friendly Eating Patterns in the New Zealand Context. Environ Health Perspect 2020;128:17007.

3. Blakely T, Cleghorn C, Mizdrak A, Waterlander W, Nghiem N, Swinburn B, Wilson N, Ni Mhurchu C. The effect of food taxes and subsidies on population health and health costs: a modelling study. Lancet Public Health 2020;5:e404-e13.

4. Cleghorn C, Wilson N, Nair N, Kvizhinadze G, Nghiem N, McLeod M, Blakely T. Health Benefits and Cost-Effectiveness From Promoting Smartphone Apps for Weight Loss: Multistate Life Table Modeling. JMIR Mhealth Uhealth 2019;7:e11118.

5. Cobiac LJ, Mizdrak A, Wilson N. Cost-effectiveness of raising alcohol excise taxes to reduce the injury burden of road traffic crashes. Inj Prev 2019;25:421-27.

6. Pearson AL, Kvizhinadze G, Wilson N, Smith M, Canfell K, Blakely T. Is expanding HPV vaccination programs to include school-aged boys likely to be value-for-money: a cost-utility analysis in a country with an existing school-girl program. BMC Infect Dis 2014;14:351.

7. Blakely T, Kvizhinadze G, Karvonen T, Pearson AL, Smith M, Wilson N. Cost-effectiveness and equity impacts of three HPV vaccination programmes for school-aged girls in New Zealand. Vaccine 2014;32:2645-56.

8. McLeod M, Kvizhinadze G, Boyd M, Barendregt J, Sarfati D, Wilson N, Blakely T. Colorectal cancer screening: How health gains and cost-effectiveness vary by ethnic group, the impact on health inequalities, and the optimal age range to screen. Cancer Epidemiol Biomarkers Prev 2017;26:1391-400.

9. Teng AM, Kvizhinadze G, Nair N, McLeod M, Wilson N, Blakely T. A screening program to test and treat for Helicobacter pylori infection: Cost-utility analysis by age, sex and ethnicity. BMC Infect Dis 2017;17:156.

10. Jaine R, Kvizhinadze G, Nair N, Blakely T. Cost-effectiveness of a low-dose computed tomography screening programme for lung cancer in New Zealand. Lung Cancer 2020;144:99-106.

11. McLeod M, Sandiford P, Kvizhinadze G, Bartholomew K, Crengle S. Impact of low-dose CT screening for lung cancer on ethnic health inequities in New Zealand: a cost-effectiveness analysis. BMJ Open 2020;10:e037145.

12. Horeweg N, Scholten ET, de Jong PA, van der Aalst CM, Weenink C, Lammers JW, Nackaerts K, Vliegenthart R, ten Haaf K, Yousaf-Khan UA, Heuvelmans MA, Thunnissen E, Oudkerk M, Mali W, de Koning HJ. Detection of lung cancer through low-dose CT screening (NELSON): a prespecified analysis of screening test performance and interval cancers. Lancet Oncol 2014;15:1342-50.

13. Collinson L, Kvizhinadze G, Nair N, McLeod M, Blakely T. Economic evaluation of single-fraction versus multiple-fraction palliative radiotherapy for painful bone metastases in breast, lung and prostate cancer. J Med Imaging Radiat Oncol 2016;60:650-60.

14. Blakely T, Cobiac LJ, Cleghorn CL, Pearson AL, van der Deen FS, Kvizhinadze G, Nghiem N, McLeod M, Wilson N. Health, health inequality, and cost impacts of annual increases in tobacco tax: Multistate life table modeling in New Zealand. PLoS Med 2015;12:e1001856.

15. Blakely T, Cobiac LJ, Cleghorn CL, Pearson AL, van der Deen FS, Kvizhinadze G, Nghiem N, McLeod M, Wilson N. Correction: Health, Health Inequality, and Cost Impacts of Annual Increases in Tobacco Tax: Multistate Life Table Modeling in New Zealand. PLoS Med 2016;13:e1002211.

16. Nghiem N, Leung W, Cleghorn C, Blakely T, Wilson N. Mass media promotion of a smartphone smoking cessation app: modelled health and cost-saving impacts. BMC Public Health 2019;19:283.

17. Nghiem N, Cleghorn CL, Leung W, Nair N, van der Deen FS, Blakely T, Wilson N. A national quitline service and its promotion in the mass media: modelling the health gain, health equity and cost-utility. Tob Control 2018;27:434–41.

18. Nghiem N, Blakely T, Cobiac LJ, Cleghorn CL, Wilson N. The health gains and cost savings of dietary salt reduction interventions, with equity and age distributional aspects. BMC Public Health 2016;16:423.

19. Cleghorn C, Blakely T, Jones A, Kvizhinadze G, Mizdrak A, Nghiem N, Ni Mhurchu C, Wilson N. Feasible diet intervention options to improve health and save costs for the New Zealand population. Burden of Disease Epidemiology, Equity and Cost-Effectiveness Programme. Wellington: Department of Public Health, University of Otago, Wellington, 2019. https://www.otago.ac.nz/wellington/departments/publichealth/research/bode3/publications/#reports. 2019.

20. Mizdrak A, Blakely T, Cleghorn CL, Cobiac LJ. Potential of active transport to improve health, reduce healthcare costs, and reduce greenhouse gas emissions: A modelling study. PLoS One 2019;14:e0219316.

21. Cleghorn C, Blakely T, Mhurchu CN, Wilson N, Neal B, Eyles H. Estimating the health benefits and cost-savings of a cap on the size of single serve sugar-sweetened beverages. Prev Med 2019;120:150-56.

22. Cleghorn CL, Wilson N, Nair N, Kvizhinadze G, Nghiem N, McLeod M, Blakely T. Health benefits and costs of weight-loss dietary counselling by nurses in primary care: a cost-effectiveness analysis. Public Health Nutr 2020;23:83-93.

23. Mizdrak A, Telfer K, Direito A, Cobiac LJ, Blakely T, Cleghorn CL, Wilson N. Health Gain, Cost Impacts, and Cost-Effectiveness of a Mass Media Campaign to Promote Smartphone Apps for Physical Activity: Modeling Study. JMIR Mhealth Uhealth 2020;8:e18014.

24. Blakely T, Collinson L, Kvizhinadze G, Nair N, Foster R, Dennett E, Sarfati D. Cancer care coordinators in stage III colon cancer: a cost-utility analysis. BMC Health Serv Res 2015;15:306.

25. Webber-Foster R, Kvizhinadze G, Rivalland G, Blakely T. Cost-effectiveness analysis of docetaxel versus weekly paclitaxel in adjuvant treatment of regional breast cancer in New Zealand. Pharmacoeconomics 2014;32:707-24.

26. Leung W, Kvizhinadze G, Nair N, Blakely T. Adjuvant Trastuzumab in HER2-Positive Early Breast Cancer by Age and Hormone Receptor Status: A Cost-Utility Analysis. PLoS Med 2016;13:e1002067.

27. Nair N, Kvizhinadze G, Blakely T. Cancer Care Coordinators to Improve Tamoxifen Persistence in Breast Cancer: How Heterogeneity in Baseline Prognosis Impacts on Cost-Effectiveness. Value Health 2016;19:936-44.

28. Sundborn G. Policy brief: a sugary drink tax for New Zealand and 10,000-strong petition snubbed by Minister of Health and National Government. N Z Med J 2017;130:114-16.

29. Jackson SE, Chester JD. Personalised cancer medicine. Int J Cancer 2015;137:262-66.

30. Haslem DS, Van Norman SB, Fulde G, Knighton AJ, Belnap T, Butler AM, Rhagunath S, Newman D, Gilbert H, Tudor BP, Lin K, Stone GR, Loughmiller DL, Mishra PJ, Srivastava R, Ford JM, Nadauld LD. A Retrospective Analysis of Precision Medicine Outcomes in Patients With Advanced Cancer Reveals Improved Progression-Free Survival Without Increased Health Care Costs. J Oncol Pract 2017;13:e108-e19.

31. PHARMAC. Making funding decisions (Fact Sheet #4). Wellington: PHARMAC. https://www.pharmac.govt.nz/assets/factsheet-04-making-funding-decisions.pdf.

Contact diana@nzma.org.nz
for the PDF of this article

Subscriber Content

The full contents of this pages only available to subscribers.

LOGINSUBSCRIBE