Developing methods for understanding mechanism in complex interventions

The general public is exposed through the media to countless claims concerning the efficacy of complex interventions such as psychological therapies for the treatment of depression or other mental health and personal problems. Health service providers are under increasing pressure to increase the availability of interventions such as counselling and psychotherapy. There is clearly a need to design and implement controlled clinical trials to test whether such interventions work.

It is equally clear that research projects need to be carried out that can inform on how the interventions work. Understanding the mechanism by which treatments affect clinical outcomes can help to develop new or improve existing treatments. For example if a treatment has a positive effect on a clinical outcome that is not due to changes in the processes targeted then the mechanism by which this beneficial effect arises can potentially be identified and the therapeutical component driving it enhanced. Similarly if a pathway is identified by which a treatment impacts negatively on a clinical outcome then steps might be undertaken to prevent it.

The proposed research will develop robust and efficient statistical methods for evaluating the mechanisms by which complex interventions operate. It will also help researchers plan trials so that they can establish not only whether but also how treatments work.

This is a proposal to develop novel methodology to support the evaluation of complex interventions in particular in mental health research. Although the recognition of the potential role of multiple therapeutic components is integral to the development of the complex intervention the mechanism by which improvements in clinical outcomes are brought about typically remains to be evaluated during the trial phase and the evaluation of putative mediators is a key research objective of an explanatory EME trial. This proposal will further develop causal inference methods for explanatory analyses to provide innovative and more powerful methodology that enables researchers to extract the most accurate information from expensive trials conducted using the latest research designs. In particular we propose to extend statistical and econometric methods to trials collecting factor and/or longitudinal data. We will also develop and implement strategies for efficiently exploiting data provided by series of trials and by trials nested within or complemented by observational data. We will address the statistical challenge of data being missing by design in such series of studies by utilising external calibration samples or data collected as part of the nested design to identify parameters describing mediation processes. We further propose to assess model-averaging as a general approach for combining instrumental variables and Baron and Kenny-type estimators to gain power. Finally, in parallel we will develop guidelines for enhanced design and measurement protocols to enable optimally informative explanatory analyses of trials of complex interventions.

Provision of Research Methodology

The project will provide research methodology to be employed by those involved in the design and conduct of intervention studies: applied statisticians, research clinicians and researchers/trialists from specific disciplines, e.g. mental health.

Evaluation model for complex interventions
The development of the capability and capacity to implement high-quality clinical research, including the evaluation of complex interventions through randomised trials, is a key priority of the NHS, the NIHR and the MRC. The evaluation of complex treatment programmes for mental illness (such as cognitive remediation therapy for psychosis) is not only a vital component of this research in its own right, but also provides a well-established model for the evaluation of complex interventions in other clinical areas. The present proposal for the development of research methodology for the optimal design, implementation and interpretation of the results of such trials is an essential component of these developments.

Research design
The project will provide guidelines to those planning complex intervention studies to ensure that design and measurement protocols are used that enable the robust and powerful evaluation of treatment effects, as well as the mechanisms by which they are induced.

NHS Policy

The project will investigate the use of nested trial designs whereby a clinical trial is embedded in routine care, and data are available on both the outcomes of the trial as well as measures collected routinely in medical practice. We will provide methods for optimally exploiting such enhanced data sources. We will make recommendations as to which measures need to be collected for subsets of patients to ensure that the mechanism by which treatments operate can be evaluated. Such recommendations will help to construct efficient (electronic) medical records systems that produce high quality essential data and thus can inform on treatment effects in naturalistic settings. Conducting trials embedded in routine care can also lead to efficiency savings and such studies hold more promise in terms of translation to patient benefit in that the additional results gained from the naturalistic component will allow generalisation to routine practice.

Health and Well-being

The general public is exposed through the media to countless claims concerning the efficacy of complex interventions such as psychological therapies. Health service providers are under increasing pressure to increase the availability of therapies. There is clearly a need to design and implement controlled clinical trials to test whether these therapies work. It is equally clear that we need to develop and implement research projects that can tell us how these therapies work and how processes might be manipulated so that therapies can be refined and improved. The project will develop research methodology which will enable researchers to understand treatment effect mechanism and in the longer term lead to better treatments.

Capacity Building

Biostatistics has long been an area of skills shortage. This project will provide training for a young statistician in skills with potential applications spanning medicine, finance, marketing and elsewhere.