Methodology State-of-the-art workshop: management of measurement reactivity in trials of interventions to improve health
Lead Research Organisation:
University of Manchester
Department Name: UNLISTED
Abstract
Measurement reactivity has been defined as being present where measurement results in changes in the people being measured. There is evidence which show that asking people to complete questionnaires can result in changes in behaviour. In addition, other types of measurement can change behaviour, specifically in medical and healthcare contexts. This workshop will examine how these changes in behaviour in response to being measured can result in bias in trials in medical and healthcare contexts.
Although there is increasing study of why measurement can affect behaviour, there remains a need to increase understanding of situations where being measured could interact with the mechanisms by which healthcare interventions are supposed to produce effects. There is a lack of consensus on the situations where one would expect measurement reactivity effects to occur, and little discussion of when one would expect these measurement reactivity effects to interact with other interventions.
In sum, although there is now good evidence that measurement is not an inert procedure, it has generally been ignored in discussions of how to reduce bias in trials. There is no agreed set of practices for conduct, reporting or analysis of measurements that allow these patterns of responding to be understood or controlled for.
Therefore, a set of guidance statements on how best to reduce bias due to measurement reactivity in studies of interventions to improve health will be produced.
Although there is increasing study of why measurement can affect behaviour, there remains a need to increase understanding of situations where being measured could interact with the mechanisms by which healthcare interventions are supposed to produce effects. There is a lack of consensus on the situations where one would expect measurement reactivity effects to occur, and little discussion of when one would expect these measurement reactivity effects to interact with other interventions.
In sum, although there is now good evidence that measurement is not an inert procedure, it has generally been ignored in discussions of how to reduce bias in trials. There is no agreed set of practices for conduct, reporting or analysis of measurements that allow these patterns of responding to be understood or controlled for.
Therefore, a set of guidance statements on how best to reduce bias due to measurement reactivity in studies of interventions to improve health will be produced.
Technical Summary
Measurement reactivity has been defined as being present where measurement results in changes in the people being measured. The clearest evidence for this comes from a number of systematic reviews which show that asking people to complete questionnaires can result in changes in behaviour. In addition, there is ample other evidence that measurement other than questionnaire effects on behaviour can produce reactivity specifically in medical and healthcare contexts.
This workshop will examine how these measurement reactivity effects can result in bias in trials in medical and healthcare contexts. The potential for bias exists where measurement differentially affects intervention and control groups. There are at least two situations where this is likely to occur: where interventions and measurements interact, and where measurement is unbalanced across randomised groups.
There is a rapidly increasing literature which aims to identify why measurement can affect behaviour. Despite this, to date there appears to have been little systematic consideration given to which other situations exist where the effects of measurement interact with the mechanisms by which interventions are supposed to produce effects.
There is a lack of consensus on the situations where one would expect measurement reactivity effects to occur, and little discussion of when one would expect these measurement reactivity effects to interact with other interventions. In sum, although there is now good evidence that measurement is not an inert procedure, it has generally been ignored in discussions of how to reduce bias in trials. There is no agreed set of practices for conduct, reporting or analysis of measurements that allow these patterns of responding to be understood or controlled for.
Therefore, a set of guidance statements on how best to reduce bias due to measurement reactivity in studies of interventions to improve health, with a particular focus on bias in RCTs, will be produced. This focus is necessary, given the huge literature on measurement reactivity that sprawls across multiple disciplines, and the central importance of trials evidence for healthcare decision-making.
To achieve the overall aim, the present research has the following objectives:
a. identify key background literature examining measurement reactivity
b. determine the scope of the guidance that would best meet stakeholder needs
c. achieve consensus on guidance statements, and produce a guidance document based on this consensus
d. disseminate key messages of guidance
This workshop will examine how these measurement reactivity effects can result in bias in trials in medical and healthcare contexts. The potential for bias exists where measurement differentially affects intervention and control groups. There are at least two situations where this is likely to occur: where interventions and measurements interact, and where measurement is unbalanced across randomised groups.
There is a rapidly increasing literature which aims to identify why measurement can affect behaviour. Despite this, to date there appears to have been little systematic consideration given to which other situations exist where the effects of measurement interact with the mechanisms by which interventions are supposed to produce effects.
There is a lack of consensus on the situations where one would expect measurement reactivity effects to occur, and little discussion of when one would expect these measurement reactivity effects to interact with other interventions. In sum, although there is now good evidence that measurement is not an inert procedure, it has generally been ignored in discussions of how to reduce bias in trials. There is no agreed set of practices for conduct, reporting or analysis of measurements that allow these patterns of responding to be understood or controlled for.
Therefore, a set of guidance statements on how best to reduce bias due to measurement reactivity in studies of interventions to improve health, with a particular focus on bias in RCTs, will be produced. This focus is necessary, given the huge literature on measurement reactivity that sprawls across multiple disciplines, and the central importance of trials evidence for healthcare decision-making.
To achieve the overall aim, the present research has the following objectives:
a. identify key background literature examining measurement reactivity
b. determine the scope of the guidance that would best meet stakeholder needs
c. achieve consensus on guidance statements, and produce a guidance document based on this consensus
d. disseminate key messages of guidance
Organisations
People |
ORCID iD |
Publications
French D
(2021)
Reducing bias in trials due to reactions to measurement: experts produced recommendations informed by evidence
in Journal of Clinical Epidemiology
French DP
(2021)
Reducing bias in trials from reactions to measurement: the MERIT study including developmental work and expert workshop.
in Health technology assessment (Winchester, England)
Miles LM
(2020)
Asking questions changes health-related behavior: an updated systematic review and meta-analysis.
in Journal of clinical epidemiology