Abstract
Introduction
Screening for earlier detection and better management of life-threatening conditions has the potential to reduce premature death and improves quality of life in the population. The evaluation of screening programmes, however, raises a number of challenges for health services research and it is not surprising that researchers have focused their attention on diverse aspects of screening and early detection. The net benefit of screening programmes depends not only on the effectiveness of screening techniques but also upon the take up of opportunities for screening amongst the at-risk population and the potential harms that may result in falsely diagnosing those who do not actually have the disease or those who have it, but will not benefit from treatment. The analysis of the cost-effectiveness of screening involves consideration of the substantial resources that are often involved, especially in national programmes targeting whole population groups. The research reported in this theme examined the evidence for the establishment of national cancer screening policies for prostate and bowel cancer; in the former case leading to a policy that recommended screening should not be undertaken, whilst in the latter a national screening programme was developed. Similarly, careful analysis of the pros and cons of screening for abdominal aortic aneurysms resulted in the implementation of a national screening programme, both in the UK and elsewhere. Researchers investigating novel ways to overcome poor uptake of screening for type 2 diabetes and pre-diabetic states, produced risk scoring tools that can be used by individuals as well as doctors, across a wide range of different settings, including on-line assessment, particularly targeted at high-risk populations. By exploring the full impact of screening, health services research has underpinned policies that save lives and make the most effective use of scarce resources.
Avoiding harm and evaluating benefit: establishing and implementing an evidence-based policy for prostate cancer screening in the UK
Summary
Prostate cancer screening is one of the most controversial healthcare topics. Prostate cancer kills over 11,000 men per annum in the UK and although screening can identify tumours, the benefits are small and uncertain. Many prostate cancers can be identified when potentially curable following screening with a prostate specific antigen (PSA) blood test and prostate biopsy, but it is not possible to identify which tumours will become aggressive or life-threatening and the vast majority will not. Screening leads to large numbers of men being diagnosed and suffering harms related to the diagnosis and treatment of prostate cancer in the context of small and uncertain benefits, hence the current UK policy not to recommend screening. Research has provided the evidence base underpinning this policy, which has avoided harm and saved substantial resources.
Research and impact
A substantial body of research evidence was assembled, beginning with a systematic review of prostate cancer diagnosis, treatment and screening literature, which concluded ‘current evidence does not support a national screening programme for prostate cancer in the UK’. 1 A major research programme followed in order to provide the required evidence to fill this gap.
First, Prostate testing for cancer and Treatment (ProtecT), a feasibility study, investigated barriers inhibiting a randomized controlled trial (RCT) of treatment. 2 Over 8500 men aged 50–69 years were recruited from general practices, and 224 men diagnosed with localized prostate cancer participated in the pilot RCT of treatments. Integrated qualitative research supported clinicians and men in accepting randomization between surgery, radiotherapy and ‘active monitoring’ (a management option developed with patients consisting of regular review and avoiding radical treatment). 2 Second, the main ProtecT RCT was launched in 2001 to evaluate the comparative effectiveness of radical surgery, radical conformal radiotherapy and active monitoring for men with localized prostate cancer. Of 111,000 men recruited with a PSA blood test, over 8500 received biopsies, and 1650 with prostate cancer were randomized between the treatment arms. A nested cohort study of men undergoing prostate biopsy investigated side-effects of biopsy, including symptoms, healthcare use 3 and psychological impact, 4 and analysis of prostate cancer mortality is ongoing. Third, the Cluster RCT of testing for Prostate Cancer (CAP) Trial randomized 573 general practices to enable a comparison between PSA testing and treatment in ProtecT (screening) and usual NHS care (control) in 415,000 UK men. Fourth, an ecological study confirmed a much lower incidence of prostate cancer in the UK compared with the USA 5 ; and last, a cohort study confirmed low rates of PSA testing in UK primary care 6 – reflecting much higher rates of screening in the USA compared with the UK.
UK policy was established in a letter from the Department of Health (DH) to all UK health authorities and clinicians in 1997, stating that: Population screening for prostate cancer, including the use of prostate specific antigen (PSA) as a screening test, should not be provided by the NHS or offered to the public until there is new evidence of an effective screening technology for prostate cancer.
7
The systematic review 1 was cited as a source of evidence, and the policy remains unchanged. The implementation of policy was supported directly by the membership of researchers on the National Screening Committee (NSC) Scientific Reference Group which launched the Prostate Cancer Risk Management Programme (PCRMP) in 2002. PCRMP issued on-line and paper documents containing information about PSA and testing and prostate cancer diagnosis and treatment, based on research evidence 1 to enable patients to make informed decisions about screening. 8 The PCRMP documents were revised in 2009 9 crediting the research as a source of evidence, and it remains the primary source of information for UK GPs and men. The effect of the policy is that the level of UK PSA-testing is low (6.2%),6,10 and there are much lower levels of incidence and treatment of localized prostate cancer compared with countries where PSA testing has been widespread since the 1980s: for example in the USA 5 , Canada, Australasia, Northern and Western Europe. 11
The research ensured that in addition to existing evidence on the potential benefits of prostate screening, there was also robust evidence on the potential harms. The cohort study of men undergoing prostate biopsy in the ProtecT study showed that 1.3% required hospital admission and 10.4% consultation with a doctor because of post-biopsy symptoms including pain, fever, and blood in urine, faeces and ejaculate. 3 Among the two-thirds of men who received a negative or inconclusive biopsy result, around 20% reported high distress persisting up to 12 weeks. 4 Concerns about the harms caused by prostate cancer diagnosis and treatment in the context of uncertain benefit, and increasing realization that high levels of PSA testing did not reflect clinical need, led to a change of policy in the USA in 2011, 12 explicitly recommending that prostate cancer screening should not be routine.
The UK NSC formally reviewed policy for prostate cancer screening in 2010 using evidence from an independent option appraisal analysis based on ProtecT trial data. In the appraisal analysis of introducing prostate cancer screening at age 50 years, annually, or every two or four years, using statistical modelling, the harms of treatment always outweighed any possible benefit of screening in each potential scenario. The clinical costs alone, without administrative costs, were estimated to be £0.6 to £1.7 billion per year. The 2010 review concluded that UK policy should remain as established in 1997 and re-iterated that any change needed to await evidence from further research. 13
Research evidence has thus underpinned UK policy and has directed attention to ensuring that known harms of testing are avoided and resources are saved.
Acknowledgements
Institutions listed as contributors to the research: University of Bristol in collaboration with University of Oxford and University of Cambridge.
References
Introduction of a national colorectal cancer screening programme
Summary
Colorectal (bowel) cancer is the third most common cancer with more than 41,000 people diagnosed with the disease each year in the UK. Approximately 16,000 people die of colorectal cancer annually. Evidence suggests that colorectal cancer screening may reduce incidence, morbidity and mortality associated with the disease. Research to evaluate the cost-effectiveness and resource implications of potential screening programmes for colorectal cancer informed the decision to launch a national colorectal cancer screening programme in England. The programme identifies individuals with less advanced colorectal cancer, and there is emerging evidence that it has led to an overall improvement in prognosis. Projections suggest that the programme is on course to reduce colorectal cancer deaths by 16%. Research has also informed subsequent re-appraisal of colorectal screening options in the NHS following new evidence and national screening policy in Ireland.
Research and impact
Research to evaluate the cost-effectiveness and resource implications of potential screening programmes for colorectal cancer has informed decisions about whether the NHS should adopt a bowel cancer screening programme and, if so, the choice of test modalities, population and frequency that should form the basis of the programme. The study included a review of existing randomized trials of alternative screening modalities, a model-based health economic evaluation and an analysis of resource implications for alternative options.
Researchers developed a health economics model to simulate the life experience of a hypothetical cohort of individuals without polyps or cancer through to the development of adenomas and malignant carcinoma and subsequent death in the general population of England. The costs, health effects and resource impact of five screening options were evaluated using this model: (a) biennial faecal occult blood testing (FOBT) for individuals aged 50–69; (b) biennial FOBT for individuals aged 60–69; (c) once-only flexible sigmoidoscopy (FSIG) for individuals aged 55; (d) once-only FSIG for individuals aged 60; and (e) once-only FSIG for individuals aged 60, followed by biennial FOBT for individuals aged 61–70. Each option was compared in terms of expected health benefits (survival/quality-adjusted life years (QALYs) gained), costs and resource implications. The economic analysis suggested that screening using FSIG with or without FOBT was likely to produce cost-savings and additional health benefits compared against no screening. However, the accompanying resource use analysis suggested that the considerable endoscopy capacity requirements associated with the FSIG screening options may make them infeasible given health service capacity constraints.1–5
This research study provided the key evidence which was reviewed by the English Bowel Cancer Advisory Group in 2004 in formulating recommendations to the Secretary of State for Health for colorectal cancer screening in England and its relationship to the subsequent policy decision is cited in advice to the NHS on bowel cancer screening. 6 The policy decision was that a national screening programme involving FOBT for individuals aged 60–69 would be launched in England. The NHS Bowel Cancer Screening Programme launched an FOBT-based programme in 2006 which is now fully rolled out across England. This policy decision resulted in a substantial service change for the NHS requiring the establishment of whole new system infrastructures (screening hubs, laboratory testing, etc.) and their integration with existing services for endoscopy. The screening programme is available to all men and women in England from the date of their 60th or 61st birthday. An extension has been rolled out to include individuals up to the age of 74 years of age. There are plans to include an additional screening using FSIG for individuals aged 55 years of age, evaluated retrospectively using the options appraisal model developed by researchers. 7
The national screening programme has in turn led to improvements in the prognosis of patients with diagnosed bowel cancer.8,9 Evidence suggests that earlier diagnosis is associated with improved survival 9 and improved health-related quality of life. 10 Analyses from RCTs and the English bowel cancer screening programme indicate that screening results in earlier diagnosis. There is a trend towards increased incidence (around 13% in the UK between 2006 and 2008) since the rollout of the programme; this reflects additional cases of preclinical cancer that would otherwise have probably have been diagnosed later at a more advanced stage. Research evidence from randomized controlled trials has shown that FOBT can reduce colorectal cancer mortality (approximately 16%) and that FSIG can reduce both incidence and mortality (23% and 31%, respectively). Statistics from Cancer Research UK indicate that the mortality rate for bowel cancer in the period 2008–2010 was 14% lower than the rate in the period 1991–1999. It is likely that a proportion of this benefit is attributable to the introduction of the national screening programme, 11 suggesting around 2500 colorectal cancer deaths may be avoided each year.
Acknowledgements
Institutions listed as contributors to the research: University of Sheffield.
References
Informing the policy and implementation of screening for abdominal aortic aneurysms
Summary
Abdominal aortic aneurysms (AAAs) affect more than 4% of British men aged 65–74 and are responsible for over 6800 deaths annually. Research evidence produced from the Multi-centre Aneurysm Screening Study (MASS) showing that screening could reduce AAA-related mortality by 42% and was cost-effective, informed the UK policy decision to introduce a national screening programme for all men aged 65 years and over. The programme offers screening to 300,000 men annually, has a 75% uptake rate and the Department of Health has estimated the health gain from a screening programme to equate to at least 130,000 QALYs over 20 years. Internationally, MASS is the most significant trial of AAA screening and provides the most robust evidence-based model of its cost-effectiveness, influencing AAA screening guidelines, policies and services, in the UK, Europe and USA.
Research and impact
Following up on initial research on the costs and cost-effectiveness of AAA screening, a multi-disciplinary collaboration established the MASS trial aimed at estimating the reduction in mortality from rupture of AAA that could be achieved by population-based screening; assessing the impact of the screening programme and treatment criteria on NHS costs, and on patients’ quality of life; and producing data to allow assessment of the potential for a national screening programme.
A population-based sample of 67,800 men aged between 65 and 74 years was recruited between 1997 and 1999, with an initial four-year follow-up period. The trial produced ‘reliable evidence of benefit’ from AAA screening aimed at reducing the 6800 annual deaths in England and Wales alone 1 and found that after only four years the cost-effectiveness of screening for AAAs was at the margin of acceptability according to the NHS thresholds then in use. Over a longer period, the cost-effectiveness would improve, with the predicted ratio at 10 years falling to around a quarter of the four-year estimate. 2 Earlier research on cost-effectiveness took the form of the first ever RCT of a screening programme for AAA. 3 Both studies were included in a 2007 Cochrane review (the MASS study contributed 67,800 of the 127,891 men) which concluded that there was significant reduction in mortality from AAA in men who undergo ultrasound screening and the cost-effectiveness may be acceptable, but needed further expert analysis.
Subsequent research developed Markov models of the cost-effectiveness of screening, using the cost data collected in the original study.4,5 The continuing follow-up work retained the aim of supplying data for a national screening programme – but increasingly also informed implementation. The 13-year final MASS trial follow-up paper was published in 2012 and reported a 42% (95% confidence interval 31 to 51) reduction in the AAA-related mortality rate from screening men aged 65–74 years. 6 Research to assist in refinement of the policy has continued with cost-effectiveness modelling of potential alternative recall intervals. 7
The assessment of the cost-effectiveness of AAA screening informed the development of UK policy. The NHS AAA screening programme began in 2009 and was fully implemented in England by Spring 2013, offering screening to around 300,000 men every year during the year they turn 65.
8
The Department of Health considered policy options for AAA screening and estimated that the net value of the option adopted was £3884 million over 20 years, valuing the health benefit at a social value of £40,000 per QALY gained.
9
Their assessment of the options was informed by the results of the MASS trial, as well as other studies in the Cochrane review and particularly relied on the economic analyses for data on costs. It noted that The main elements of the cost analysis are therefore based on the outputs and subsequent analysis from MASS … The unit costs for screening and elective and emergency surgery operations are based on MASS trials … An alternative cost base … was also considered. However, the MASS unit costs are more comprehensive and reliable, and are based on a detailed bottom-up costing, taking into account patient-specific costs.
10
The MASS trial also had extensive international impact on advisory committees, guidelines, and policies and helped generate improvements in both publicly and privately funded healthcare services. A 2009 practice guideline from the US Society for Vascular Surgery drew on the same four studies as the Cochrane review in which the MASS study had the most influence. The guideline recommended ultrasound screening for AAA for all men at or older than age 65, describing the level of recommendation as ‘strong’ and the quality of evidence as ‘high’. 11 AAA screening is now widely available in the USA. Many of the policies and practices in the period from 2008 drew on a key 2005 evidence synthesis and Recommendation Statement from the US Preventive Services Task Force in its public advisory role. The MASS study’s importance was highlighted in both the recommendation and the synthesis, which have remained in place, with particular attention paid to the quality of the evidence. 12 That review formed the basis for the legislation under which Medicare has offered AAA screening from 2008.
The development of screening policies in Sweden has been informed by the evidence from the MASS trial, 13 as has the European Society for Vascular Surgery guidelines supporting population screening of older men in order to reduce aneurysm-related mortality by almost half. 14
Acknowledgements
Institutions listed as contributors to the research: Brunel University in collaboration with the University of Cambridge and the United Medical and Dental Schools of Guy’s and St Thomas’ Hospitals.
References
Pre-diabetes and type 2 diabetes: risk-assessment tools for early detection and prevention
Summary
Around 2.5 million people in the UK have type 2 diabetes (T2DM), with many more in a prediabetic state. Both conditions are hard to detect and frequently remain undiagnosed and untreated for years. The cost burden to the NHS of eventual treatment is estimated at £10 billion, 80% of which is spent on avoidable complications. Targeted diabetes prevention programmes could aid in the reduction of prevalence and associated costs. Researchers developed two risk scores to detect these conditions, both suitable for use with an ethnically diverse UK population: a self-assessment questionnaire and a general practice database tool. Recommended by NICE, they have been used successfully in varied settings. Over a two-year period, around 260,000 people completed the self-assessment score online and more than 40,000 through other means.
Research and impact
Early detection is important in the control of diabetes. T2DM is one of the most common long-term conditions globally; it is also non-communicable and potentially preventable. In the UK, around 10% of the adult population over 45 years has T2DM and diabetes prevalence is increasing steadily due to an ageing and more ethnically diverse population, as well as high obesity levels. T2DM is preceded by a high-risk prediabetic state referred to as impaired glucose regulation (IGR), this high-risk group is also known as prediabetes and non-diabetic hyperglycaemia. Population-based screening studies have shown that around 20% of the adult population over 45 years has undiagnosed T2DM or IGR, and this prevalence is significantly higher (up to six times) in those from black and minority ethnic groups. However, despite evidence to show that timely treatment can delay or prevent the development of complications, 1 uptake in screening programmes is generally low and lower still in black and minority ethnic populations where risk is higher. Research has highlighted a number of barriers to the uptake of screening: the time commitment involved in attending a long appointment during working hours, low self-perception of risk, difficulties with arranging general practice appointments, and dislike of oral glucose tolerance tests. 2 The findings highlighted the need for novel screening strategies.
Researchers developed two risk scores suitable for an ethnically diverse population, based on initial research that offered screening within 20 general practices to 30,950 randomly selected people from Leicestershire, who had not been previously diagnosed with T2DM. 3 Of these, 6390 took up the offer and a fifth were identified as either having undiagnosed T2DM or at high risk of developing the disease. A comprehensive set of data – age, sex, BMI, waist circumference, family history of diabetes, history of high blood pressure and ethnicity – as well as oral glucose tolerance tests, was collected from all those attending. The data were used to develop two scores to detect undiagnosed IGR or T2DM in an ethnically diverse population.4,5 One score is based on factors known to the patient and other is designed to enable general practices to identify those at highest risk in their populations, using routinely stored data (minus waist circumference). These were the first scores that have been specifically developed or validated in a UK multi-ethnic population and they were externally validated using data from a second screening study. 6
The tools have been successfully used to identify those at risk in a number of clinical studies. The Let’s Prevent Diabetes study used the score to rank people by risk of IGR/T2DM within general practices, and those with the highest 10% of scores were then invited for second-stage screening and a quarter of those screened were found to have either IGR or T2DM. 7 Similar results were seen for a study into the Walking Away from Diabetes programme, a 3 hour course which offers participants the opportunity to explore their risk of developing diabetes and to identify the changes they need to make to remain healthy. The self-assessment score is being used in two Collaboration for Leadership in Applied Health Research and Care (CLAHRC) early detection studies, ATTEND and PRISM, which have screened over 3000 participants to date.
Both tools are now recommended by NICE for identifying those at high risk of T2DM,8,9 and researchers were invited to lead on developing the content and structure of the Handbook of Vascular Risk Assessment, Risk Reduction and Risk Management (both the original 2008 version and updated 2012 version), 10 which is widely used within the Department of Health and across the NHS. The handbook was also a major influence in the development of the Department of Health's Best Practice Guidance. 11 Modelling work by NICE showed that utilizing risk scores in a screening programme is cost-effective and is likely to be cost-saving in those from black and minority ethnic backgrounds and research estimates that incorporating risk scores reduces the cost per case detected from £350 to around £200. 12
The self-assessment tool is being used in a wide variety of contexts. Clinical commissioning groups around the country are using it at screening events. It is also used to target people for the NHS Health Check, a programme to prevent heart disease, stroke, diabetes and kidney disease. It identifies those at high risk, who can then be given support and lifestyle interventions to reduce their risk and prevent onset of these conditions. Diabetes UK has used the self-assessment tool at local road shows since 2011, referring on those found to be at high-risk (3700 people in 2012). In a survey conducted in 2011, of those who were in these higher risk categories, 69% had been to their general practice or intended to go. Evaluation of the use of the self-assessment score within Diabetes UK activities showed that after being risk assessed: 41% had started to eat more healthily and a further 44% intended to; 33% had increased their physical activity levels and a further 43% intended to; and 44% of those referred to their GP had been to their GP to seek a test. 13 Diabetes UK has also recruited ‘community champions’ to raise awareness within black and minority ethnic communities, some of whom have been trained to undertake the risk assessments; and the test has been used with 200 people in three faith centres in Leicester, by interpreters using iPads and is being adapted for use by non-English speaking Guajarati speakers.
It has also been made available in pharmacies in England and Wales, as part of specific diabetes campaigns, as well as featuring on the websites of Diabetes UK, Boots and Tesco and is to be added to the Royal College of General Practitioners website. Around 300,000 people have completed the self-assessment tools through the Diabetes UK website, general practices, road shows, faith centres, Boots and Tesco pharmacies and PCT-organized screening events. If T2DM could be prevented in just 3% of these people as a result, this could produce a gross saving for the NHS of £40 m per year over four years.
Researchers are involved in the development of risk scores for use in Portugal and Spain, as well as a global risk score to be used in developing countries.
Acknowledgements
Institutions listed as contributors to the research: University of Leicester.