Canada Spends Logo

Explore our Budget 2026 analysis

Grant: Operating Grant

McGill University

$188,249.00

Methods to measure (and measures of) the (actual) mortality reductions produced by cancer screening.

FY 2010-11

Summary
We spend a lot on screening. We can count the dollars and document the harms to individuals. Sadly, despite many long costly trials involving large numbers of people, policy makers, funders and the public do not confidently know how much these programs have (or would) cut the numbers of breast, colon, prostate, and lung cancer deaths. One reason is the metric used, which averages over the full (but arbitrary) period of follow-up. This metric is suitable for interventions that result in virtually immediate and long-lasting reductions in sickness and death rates, such as adult circumcision, HPV/polio vaccine, or screens for abdominal aneurysms. However, it mis-measures the delayed effect of screening. The benefit of screening seen in 2011 is not due to screens done in 2011, but earlier ones. Returns on 2011 screens will emerge in mid, but mostly the late 2010s. The usual metric ignores critical timing information and badly underestimates the impact of a screening program. Naively pooling results from several trials further dilutes it. We will extend the method we developed to re-analyze data from the European Prostate Cancer Screening trial. Our method estimates the yearly benefit of a screening program by summing returns on past financial investments. Using this method, we estimated that the reduction in mortality was 50% beyond year 7, much lower than the 20% reduction first reported. We will refine our method to handle data from trials with varying schedules, thus making proper use of all available data. We will test the methods on simulated situations where we know the correct answer. We will also apply it to data from all completed trials on breast, colon and lung cancer and report the results in a meaningful way. We will also develop methods for planning the size of new screening trials. The methods described in this proposal will give policy makers and funders more accurate and relevant data on how effective cancer screening programs are/could be.
Program
Operating Grant
Location
Quebec, Canada
Recipient Type
Academia
Type
Grant
Source
CIHR Funding Decisions Database

As published

Id
225724
Name
Hanley James A
Region
Québec
Country
Canada
Orgtype
Institution Paid; Research Institution
Instname2
Health Services and Policy Research
Theme2
Health systems / services
Prcname2
Public, Community & Population Health
Projecttitle
Methods to measure (and measures of) the (actual) mortality reductions produced by cancer screening.
Programtype2
Operating Grants
Orgnameinp2
McGill University
Cihramount2
$188,249
Cihrequipment2
$0
Programname2
Operating Grant
Keyworddelim
Biostatistics; Cancer; Data Extraction; Likelihood Methods To Combine Data; New Statistical Methods; Parameter-Estimation; Sample Sizes; Screening
Abstract
We spend a lot on screening. We can count the dollars and document the harms to individuals. Sadly, despite many long costly trials involving large numbers of people, policy makers, funders and the public do not confidently know how much these programs have (or would) cut the numbers of breast, colon, prostate, and lung cancer deaths. One reason is the metric used, which averages over the full (but arbitrary) period of follow-up. This metric is suitable for interventions that result in virtually immediate and long-lasting reductions in sickness and death rates, such as adult circumcision, HPV/polio vaccine, or screens for abdominal aneurysms. However, it mis-measures the delayed effect of screening. The benefit of screening seen in 2011 is not due to screens done in 2011, but earlier ones. Returns on 2011 screens will emerge in mid, but mostly the late 2010s. The usual metric ignores critical timing information and badly underestimates the impact of a screening program. Naively pooling results from several trials further dilutes it. We will extend the method we developed to re-analyze data from the European Prostate Cancer Screening trial. Our method estimates the yearly benefit of a screening program by summing returns on past financial investments. Using this method, we estimated that the reduction in mortality was 50% beyond year 7, much lower than the 20% reduction first reported. We will refine our method to handle data from trials with varying schedules, thus making proper use of all available data. We will test the methods on simulated situations where we know the correct answer. We will also apply it to data from all completed trials on breast, colon and lung cancer and report the results in a meaningful way. We will also develop methods for planning the size of new screening trials. The methods described in this proposal will give policy makers and funders more accurate and relevant data on how effective cancer screening programs are/could be.
Competitiondate
201103
Approvedterm2
3 yrs 0 mth
Pinamesdelim
Hanley, James A
Approvedterm
3 ans 0 mois
Cihrcontribution2
~ $188,249~ $0~