Fifteen years of bowel cancer screening policy in Australia: putting evidence into practice?

Over 4000 Australians die from bowel cancer each year, making it the second biggest cause of cancer death in Australian men, after prostate cancer, and in Australian women, after breast cancer.1 Meta-analysis of randomised controlled trial (RCT) evidence has shown a reduction in relative risk of dying from bowel cancer of 25% for those who attended at least one round of faecal occult blood test (FOBT) screening.2 In addition, FOBT screening has the potential to reduce morbidity from bowel cancer through earlier diagnosis and treatment, and prevent bowel cancer by the removal of precancerous lesions, identified through colonoscopic follow-up of positive FOBTs.3 FOBT screening is the only mode of bowel cancer screening with RCT evidence of efficacy.

Here, we describe the history of bowel cancer screening policy development and implementation in Australia, based on semi-structured interviews with key informants and document analysis (Box 1). Research evidence appears to have played a major role in the development of bowel cancer screening policy proposals. However, policy decisions about how best to implement the program, given considerable external constraints, have not been based on evidence and reflect a lack of understanding of the importance of research evidence in population screening programs, where some people will inevitably be harmed and a few will benefit.

The development of bowel cancer screening policy proposals from the late 1980s to October 2005 was largely evidence based (Box 2). Clinicians made a case for considering FOBT screening, after publication of the first RCT evidence,7 and the federal government agreed to establish an expert group (the Australian Health Technology Advisory Committee [AHTAC] Working Party on Colorectal Cancer Screening)4 to specifically review the benefits, risks and costs of bowel cancer screening in Australia. This review comprised a detailed examination of literature, including evidence from recently published RCTs7,12,13 and cost-effectiveness analysis.11

After the AHTAC’s report was published in December 1997,4 the federal government allocated funding in the 2000–01 budget for a pilot program of FOBT screening, which ran from November 2002 to June 2004. Invitations were sent to 57 000 people in three states, with households randomly allocated to receive one of two immunochemical FOBT kits. The Bowel Cancer Screening Pilot Implementation Committee commissioned a range of qualitative and quantitative studies, including an unpublished literature review and cost-effectiveness analysis, and concluded that an FOBT screening program would be feasible, acceptable and cost-effective in the Australian setting.5 The final report on the pilot program included a proposed framework for implementing a national program with biennial screening using an immunochemical FOBT for the initial target age group of 55–74-year-olds and a review of the age range once the program was fully operational.5

The methodical approach to examining evidence on bowel cancer screening was out of kilter with the 3-year electoral cycle in Australia and the more practical demands on policymakers. Despite the expert review of evidence by the Bowel Cancer Screening Pilot Implementation Committee, the final report was not published until October 2005 (16 months after the pilot program was completed) and so was not publicly available in time to inform policy proposals on this issue. Advisers of then Health Minister Tony Abbott had already identified bowel cancer screening as one of a suite of cancer-related options for inclusion in their 2004 election campaign document Strengthening cancer care,17 published before the election held on 9 October. The pressure of an upcoming federal election left little time for considered reflection of the evidence by the politicians responsible for making the decisions, and the government’s campaign promises reflected this lack of evidence.

In Strengthening cancer care, the Howard government pledged to “work to phase in a national, coordinated, population-based screening programme based on the pilots, with the ultimate aim that, by 2008, every Australian aged over 55 and Indigenous Australians aged over 45 will be screened for bowel cancer at least every two years”.17 This election commitment indicates that there was sufficient political will to fund a national program of bowel cancer screening, but it also reflects the lack of understanding of screening issues by those making the commitment. For example, it set no upper age limit on those to be screened, advocated an earlier starting age for Indigenous Australians, unrealistically assumed that a full roll-out was possible within a short time frame, and proposed a budget of only $25.5 million over 4 years17 — a small fraction of the likely estimated cost,17 given the Rudd government’s subsequent allocation of $29 million per year for just three age groups.18

Abbott’s policy advisers were not the only ones to get the figures so drastically wrong. Once a federal election has been called in Australia, a set of caretaker conventions limits the access that political parties have to the Australian Public Service.19 However, both the Prime Minister and the Leader of the Opposition have the right, under the Charter of Budget Honesty Act 1998 (Cwlth), to request costings of publicly announced election commitments. The Department of Finance and Administration, at the request of the then Prime Minister, costed the entire Strengthening cancer care package at $130.1 million over 4 years, with $34.5 million allocated to “cancer prevention (bowel cancer, smoking and skin cancer)” over 3 years — again a major underestimate of probable costs.20 Ironically, costings for the proposed national bowel cancer screening program had already been prepared by M-TAG as part of the pilot process. Although this report — Cost-effectiveness evaluation of a national bowel cancer screening program — was not published, bureaucrats working in the Department of Health and Ageing would have known about the costings, and the Secretary to the Department of Finance and Administration would have been within their rights to request the information. Yet, there appears to have been no sharing of these costings between departments.

All new major federal policy proposals are vetted by the Expenditure Review Committee, a subcommittee of Cabinet, which includes senior representatives from Treasury, the Department of Finance and Administration and the Department of the Prime Minister and Cabinet. Despite 2004 being a time of economic surplus, once the inaccurate Department of Finance and Administration costings had been approved by the Expenditure Review Committee, and the rest of the budget had been allocated to other programs, the institutional constraints of this system made subsequent budgetary allocations of the magnitude required virtually impossible. The opportunity for a fully funded national bowel cancer screening program, as outlined in the election promises, had passed.

In the May 2006 federal budget, $35.6 million over 3 years was allocated to fund the first phase of the National Bowel Cancer Screening Program.16 This was around $10 million more than the figure costed in the election campaign document, but fell far short of the real costs of a full program and was a major constraint on program implementation options. For example, there was no federal funding for the states to cover the costs of follow-up colonoscopies for those with positive FOBTs performed in the public sector, so additional colonoscopies were to be incorporated into the “usual care” model and the number of participants who could be screened was limited. From August 2006 to June 2008, the Howard government offered screening to those turning 55 and 65 years of age and to those who were invited to participate in the pilot program but declined, and offered rescreening to those who were screened in the pilot.21

Official justification for the choice of these age groups was, and continues to be, based on the need to manage the capacity of health services given the expected increased demand for colonoscopy and treatment services.21,22 Although this is a valid argument for a staged roll-out, it does not justify the choice of these two particular age cohorts, which appears to have been based primarily on cost — we were unable to find any published cost-effectiveness analysis data on screening at particular ages, rather than age ranges. Existing international evidence and practice suggests that screening people older than 60 years is likely to be the most cost-effective strategy, as more cancers and precancerous lesions would be detected in an older population, and 60-year-old men have been shown to have much higher cumulative 10-year mortality benefits than men aged 40 or 50 years.2 According to recent findings from the United Kingdom, 61% of bowel cancers occur in those aged 70 years or over.23 The English program screens those between 60 and 69 years of age, and plans have been made to extend this age range to 75 by 2010. A comparison of Australian, English and other national FOBT programs is shown in Box 3. Economic modelling of a potential bowel cancer screening program in Australia, published in 2004, showed that extending the age range for screening upwards to the 75-years-and-older age group was likely to be cost-effective, but extending it downwards to the 50–54-year age group was significantly less likely to be cost-effective.29 However, no-one over 65 years is included in the Australian national program, and this mismatch between the evidence and choice of ages was compounded by the Rudd government’s announcement in the May 2008 budget of $87 million over 3 years to continue the existing program and extend screening to 50-year-olds from 1 July 2008.18

Commencing bowel cancer screening at age 50 years is in line with the most recent National Health and Medical Research Council (NHMRC)-approved guidelines,30 and would be a welcome policy decision if resources were less scarce and could allow for biennial screening (as the guidelines stipulate). However, Health Minister Nicola Roxon has since made it clear that there will be no biennial screening under the second phase of the national program, which runs until June 2011.31 This means that 50-year-olds will be rescreened in 5 years, 55-year-olds will be rescreened in 10 years, and 65-year-olds will not be rescreened at all. In February 2009, perhaps in recognition of the lack of biennial screening, the government placed an FOBT kit on the Medicare Benefits Schedule, so some reimbursement from the government is now possible.32 Results from kits used by individuals outside the national program will not be incorporated into the National Bowel Cancer Screening Program Register, and so will not contribute to ongoing data collection on how many people have been screened and the outcomes.

The latest report on the National Bowel Cancer Screening Program, which provides information for 2008, revealed a drop in overall participation rates to an estimated 39.3%, from the 45.4% achieved in the pilot program.33 Although some of this drop might be explained by the late inclusion, from July 2008, of 50-year-olds — who are less likely to participate33 (another reason to possibly focus on an older age range) — it also reflects continuing problems with obtaining follow-up data, an issue that is not confined to FOBT results. Although 64.5% of those with a positive FOBT result were recorded as having colonoscopic follow-up, this figure may underestimate the true rate because of under-reporting by clinicians and the use of positive FOBT results as a substitute denominator for colonoscopy referrals (a consequence of the low return rate of colonoscopy report forms).33 In May 2009, the roll-out of the program had to be temporarily suspended following a modification by the manufacturers to the buffer solution used in the FOBT kit, which increased the likelihood of a false-negative result. Routine evaluation of test results showed a noticeable decline in the rate of positive results from the modified kits that were distributed between 1 December 2008 and 8 May 2009. The Department of Health and Ageing estimated that around 600 000 replacement kits had to be issued, including 108 000 kits sent to people for whom negative results were returned during this period and who needed to repeat the FOBT.34 As a result, the roll-out of the program has been significantly delayed.

It may seem a reasonable policy option to start a public health program by doing what you can afford first and hoping to expand it later. A staged roll-out is a sensible approach, provided the selection of age cohorts and the frequency of screening are based on evidence. Given that the final report of the Australian pilot program concluded that FOBT screening was feasible, acceptable and cost-effective, based on biennial screening of 50–74-year-olds,5 one must question whether the current program — one-off screening of individuals aged 50, 55 and 65 years — is justifiable. It seems likely to result in fewer benefits compared with other more evidence-based age and interval options that are possible within the budgetary constraints (such as the English model).

A related concern is that if the benefits side of the equation is reduced, then the benefits may no longer outweigh the downsides of bowel cancer screening, which include anxiety due to false positives, inappropriate reassurance due to false negatives, morbidity and the remote possibility of mortality due to follow-up colonoscopies. The absolute mortality risk reduction figures for bowel cancer screening are modest — for example, it is estimated that biennial screening over a 10-year period saves one life for every 1000 men aged 55–74 years who are at average risk, and one life for every 1000 women aged 65–74 years who are at average risk.35 On the flip side, many more will not benefit and may be harmed.36,37

Finally, as there are no published data on the cost-effectiveness of one-off screening of individuals aged 50, 55 and 65 years, we do not know how cost-effective this strategy is. An independent cost-effectiveness analysis by the Cancer Institute NSW,38 published in August 2008 (around the same time as Roxon’s announcement that there would be no biennial screening under the current program), stated that the National Bowel Cancer Screening Program represented value for money for the Australian health care system, based on a model in which people turning 55 or 65 years of age were first invited to participate, and invitations to rescreen were repeated biennially thereafter until 75 years. Other eligibility scenarios, with first invitation at 45 and 50 years of age, were also shown to be cost-effective, provided that screening was offered biennially and over a specified age range.38

Although there is solid RCT evidence that bowel cancer screening can contribute to a reduction in the 80 deaths a week from this disease in Australia, policy decisions about implementation of the Australian program to date have not been based on the full extent of available research evidence, but on selected parts of the evidence. Political, financial and institutional constraints combined to shape and limit the National Bowel Cancer Screening Program. Given these constraints, the most useful basis for making evidence-based decisions about who to screen and how often to screen is age-specific cost-effectiveness analysis of Australian cost data. To our knowledge, no such analysis has been published. Fifteen years on from the government’s first involvement in bowel cancer screening, policies concerning age cohorts and screening intervals seem to be based on what the government deems it can afford, and not on firm evidence of what is required for a screening program to obtain worthwhile mortality and morbidity benefits. To realise the expected public health gains from a national bowel cancer screening program, the government needs to return this important public health policy to its evidence-based beginnings. An age-specific cost-effectiveness analysis would be a good start.