CUP RFA: Development of systematic review methodology for mechanistic studies

(Co Principal Investigator is Professor Richard Martin)

  • Topic: Cancer-related outcomes
  • Institution: University of Bristol
  • Country: United Kingdom
  • Status: Completed

Scientific abstract

(View plain language abstract)

Background

Systematic reviews are acknowledged as the most rigorous way to synthesise data which have addressed a common question. Methods for conducting and reporting rigorous systematic reviews of epidemiological studies have been developed, but remain lacking for mechanistic studies. Given the plethora of published mechanistic studies in the field of diet and cancer, some way of judging the quality and reproducibility of these studies is needed.

Objectives & hypothesis

The aim of this project is to draw on the expertise (in systematic reviews of epidemiological studies and in experimental studies of cancer) in our group in order to develop and publish a template for carrying out rigorous systematic reviews of mechanistic studies. The key objectives which will be addressed are: to design a comprehensive search strategy that will identify the diverse study types of relevance, to determine inclusion/exclusion criteria based on a hierarchy of evidence and quality control criteria which will arise from this project, to identify and quantify publication bias, which is likely to be substantial, and to test our methodology in a feasibility study designed to examine the hypothesis that milk intake increases prostate cancer risk.

In the fiscal year 2013/14 some additional work was approved for this grant. This additional work, which fitted within the first objective of the grant (to design a comprehensive search strategy that will identify the diverse study types of relevance) is to develop a software tool to automate the early stages of the search. The initial stage search is inclusive (hypothesis-free) and likely to uncover a large number of studies. The automated tool will allow visualisation and quantification of the potential mechanisms for a given exposure-outcome association by the quantity of evidence underlying each step in the mechanistic pathway.

Setting & Methods

This study will be based in the School of Social and Community Medicine at the University of Bristol (one of the UK's leading centres for research and teaching in population health sciences), but with contributions from scientists in other Schools within the University of Bristol and from the University of Cambridge and the International Agency for Cancer Research. We have selected a large multi-disciplinary team to carry-out this project with expertise in informatics, statistics, cancer biology, epidemiology, genetic epidemiology, nutrition and systematic reviews. In order to achieve our objectives we will hold a series of 4 one day workshops in which we will brainstorm the expertise within the group, which spans many different cancer sites and disciplines of cancer research. The workshops will be a mixture of presentations, small-group discussions and open discussions. Between meetings we will carry-out literature reviews, assess our methodology and ask each member of the team to comment on draft documents arising from the meetings and related research, and in doing this we will arrive at a consensus template for systematic reviews of mechanistic studies of diet and cancer. We will also carry-out a systematic review to summarize the evidence from mechanistic studies underlying the associations between milk and prostate cancer. This will feed into and will be carried out in parallel with developing the systematic review methodology above.

For the automation tool a piece of software will be developed to count co-occurrence of exposure/mechanism and mechanism/outcome terms in abstracts, generate a weighted score of importance of mechanisms and generate a graphical overview of the quantity of published evidence for each mechanism.

Potential Impact

Developing a protocol to carry-out systematic reviews of diet and cancer will allow researchers to assimilate the overall evidence in order to test any hypothesis in this area. In doing this it may help to determine whether there is a causal link between an exposure and cancer and could also highlight areas where further research is necessary. This will ultimately contribute towards preventing cancer via dietary advice and interventions.

The planned standalone automated tool has the potential to reduce time, workload and the expertise involved in implementation of the mechanism discovery process. This will make the whole methodology more efficient and accessible to a wider range of users.


Plain language abstract

Background

The WCRF International Continuous Update Project (CUP) was established to keep the evidence and Recommendations for Cancer Prevention from the 2007 Second Expert Report current. The team at Imperial College London, using the most rigorous methods possible, is collating evidence from epidemiological studies. In order to draw conclusions and make recommendations the CUP Expert Panel requires information from both epidemiological (human) and mechanistic (experimental studies in animals, humans and cells) research to help explain the links between diet, physical activity and body fatness, and cancer that are seen in epidemiological studies.

Rationale

Information from mechanistic studies was collected as part of the Second Expert Report in the form of narrative reviews (not systematic reviews). Increasingly, researchers are now highlighting the importance of systematic reviews of these types of studies but there is currently no existing systematic method similar to that followed for epidemiological studies for the CUP. The inclusion of systematically reviewed mechanistic data is critical to the future development of the CUP, and so WCRF/AICR, with the support of the CUP Expert Panel, has decided to develop a more thorough review process for the review of research on mechanisms. This is being done through a special CUP mechanisms RFA call.

Aims & Objectives

A special group, chaired by Dr Stephen Hursting (University of Texas, US), was commissioned to develop some guidelines on how a review of mechanisms could be approached. These guidelines now need to be further developed by an external group and tested for their feasibility.

A series of workshops with experts (including CUP Secretariat) is planned by Dr Sarah Lewis in order to further refine the guidelines. The feasibility test will involve the team of researchers conducting a test systematic review on milk and prostate cancer using the new instructions. The Second Expert Report found limited evidence that milk increases the risk of prostate cancer and so this is an important area where more data is required.

When complete this new methodology will be applied to all mechanistic reviews required for the CUP and will ensure cutting edge methodology is used for all aspects of the project. Systematic reviews of both the epidemiological and mechanistic studies will form the basis of the Recommendations for Cancer Prevention to be published in 2017.

In the fiscal year 2013/14 some additional work was approved for this grant. As part of the grant objectives the team will also develop a software tool to automate the early stages of the search process. The tool will also represent visually the potential mechanisms identified by the search and show the number of studies for each possible mechanism, which will help narrow down the search.

Potential Impact

This ground-breaking research will ensure the CUP remains at the forefront of cutting edge science on mechanisms. It will ensure that the Recommendations for Cancer Prevention are based on the most robust science, taking into account information from both epidemiology and mechanistic research. The impact of this work will extend beyond the CUP as this new methodology could become the mainstream approach to reviewing mechanistic studies by researchers around the world and may inform the direction of future research in the area of diet, physical activity, body fatness and cancer.

The planned automated software tool has the potential to reduce time, workload and the expertise involved in implementation of the mechanism discovery process. This will make the whole methodology more efficient and accessible to a wider range of users.