
Home Review results by therapeutic healthcare area About evidence-based healthcare The Cochrane Collaboration About health consumer participation Terminology Process
Some of the language used in clinical trials and systematic reviews Introduction Looking into clinical trials, to participate or to look at the findings - here are some of the scientific language you are likely to come across. Terminology An interventional clinical trial includes all prospective research studies that assign humans or groups of humans to one or more interventions for the purpose of evaluating their effects on health outcomes. Clinical trials are important sources of scientific evidence on the safety and effectiveness of health interventions. Access to information about ongoing, completed and published clinical trials is essential for appropriate decision-making. Researchers, research funders, policy-makers, medical practitioners, patients and the general public need such information, to help guide research or to make treatment decisions (www.who.int/ictrp). Evidence-based health care is a combination of current best research evidence, the expertise of the healthcare provider, and patient values - making decisions about health care for and with individuals, families and communities. The ‘gold standard’ for producing evidence on healthcare interventions is the double-blind, randomised clinical trial (RCT). An RCT randomises people to different treatments, as a way of trying to make the different groups as similar as possible. The intervention that is being tested is compared to a non-active (placebo), a known treatment or usual care. The people who give the treatment and assess the results of giving the treatments do not know who is receiving what intervention (double blinded). The larger the trial the more the influence of things we do not know about, including biological variations and behaviours from one person to another, can be reduced and the assessment of the effect of the treatment more accurate. What an RCT does is measure the likely efficacy of a treatment – can the intervention be shown to work in this situation with carefully selected participants, a defined healthcare setting and careful follow up? The evidence from an RCT needs to be translated into everyday use in a social context and in busy healthcare systems. Furthermore, RCTs are usually short-term, so they tell us little about long-term benefits or harms of a healthcare intervention. A systematic review of randomised controlled trials gathers individual trials and summarises them in a single review. A systematic review uses set methods to identify this best evidence and, where possible, pools the results in an overview or statistical meta-analysis. Studies are only included in the review if they meet specific, pre-specified criteria defined to minimise introducing biases and influencing the findings of both the individual studies and the review. Randomly assigning trial participants to an intervention group in a way that no-one is aware what group the person is in is an important part of these criteria. Many trials have insufficient numbers of participants to be able to distinguish intervention effects - because of natural variations between people and their responses to treatments (they are underpowered). Some of the terms used in systematic reviews Abstract: A brief, clearly structured summary of the main features and results of a study. Adverse effect: An unfavorable, or even harmful, effect or event that occurs as a result of a healthcare intervention. Allocation concealment: This is a process used in studies that involve different groups receiving different interventions or treatment. Ideally, allocation to the different groups is done in such a way that the participants, and the health care providers, do not know which intervention the participant is to receive. The intention is to avoid bias during the allocation process so that the intervention and control groups are as similar as possible. Bias: Any factor, recognised or not, that distorts the findings of a study. Blinding: A process where the participant, healthcare provider, or person assessing the outcomes (double blinding) are prevented from knowing which intervention the participant is receiving. Characteristics of a study: Identifying features of a healthcare study that include the design of the study, participants, setting, interventions used, the outcomes monitored and follow-up of the people who were initially assigned to the study. Confounder: A factor other than the intervention under investigation that could influence the study results and is sometimes difficult to identify. The factor is separately related to the outcome such that it distorts the relationship between intervention and outcome. Controls or control group: The group in a study that does not receive the intervention that is being investigated. Controls may receive no intervention, a placebo, a commonly used treatment for that condition, or only a part of an intervention. Controlled trial: An intervention trial in which a group given the intervention under study is compared to a control group which does not receive the intervention. Effectiveness: The ability of an intervention to produce a desired effect in an individual. Efficacy: The ability of an intervention to produce a desired effect in a controlled study. Eligibility criteria: The criteria that are clearly set out beforehand to define whether a potential participant is eligible to enter a study. In the Cochrane context, this term refers to the criteria that are clearly set out beforehand to define whether a study meets the requirements for inclusion in a systematic review. Epidemiology: The discipline which studies the causes, distribution, and contributing factors to healthcare problems in population or community groups. Grey literature: Reports of studies that do not appear in the peer-reviewed literature. These reports are often not easily accessible and may not be in the format of papers in the general medical literature. Intervention: A treatment, procedure or program of health care that has the potential to change the course of events of a healthcare condition. Meta-analysis: A way of analysing information or data from a number of different studies to determine an average, or common, effect. It is aimed at improving the precision of the available data by looking at a greater number of people, including from different populations. Morbidity: Illness or harm. Objective(s): The clearly stated healthcare question(s) that a study or review sets out to answer. Outcome: A measure of the effect of an intervention that is relevant to health care. Participant(s): Individual(s) who take part in a study. Placebo: An intervention that to all intents and purposes appears to be the same as that which is being assessed but which does not have the active component being assessed. Placebo effect: A change due to an expectation that the treatment or procedure will have an effect rather than the treatment or procedure itself. Population: This term can refer to the participants involved in a healthcare study; it can also refer to a general population of people. Protocol: An outline of a healthcare question, its relevance, and how reviewers intend to go about answering the question. The process of developing a protocol is aimed at improving the quality of a future review and minimising bias by deciding beforehand the methods, selection criteria and outcomes to be assessed. A protocol is refereed (peer reviewed) and published in The Cochrane Library to invite feedback as to the usefulness of the question and the process as outlined. Quasi-randomised controlled trial: This term is loosely used to refer to randomised controlled trials where the method of allocation to the different groups is not sufficiently rigorous to ensure allocation concealment. These methods include alternate patients and allocating patients by birth dates. Randomisation: A method of allocation based on chance such that individual allocation cannot be anticipated and statistical theory can be used for analysis. It is a way of ensuring groups are as similar as possible at the beginning of a study. Randomised controlled trial: A study design in which individuals are assigned, by special randomisation techniques, to two or more groups where one group receives the intervention under investigation and the other(s) receives no treatment, a placebo, or a standard intervention. Search strategy: The way a search is carried out to identify publications or reports that are relevant to a review process. Bias can be minimised and the quality of the review improved by the thoroughness of the search. Standard treatment: The treatment, procedure or intervention that is normally or conventionally given for a condition. Statistical analysis: The process of examining collected numerical measurements in order to determine if they are a fair representation of a situation, what value or significance can be given to them, and consequently the size and numerical significance of an intervention effect. Study: A systematic, well-planned investigation of a healthcare problem. Subgroup: When participants of a study are further divided according to factors other than the intervention received eg age, sex, severity of disease or physical condition, dose of intervention or quality of study. Some statistical terms Absolute risk reduction (ARR) or risk difference: The difference in the rate of outcomes between the control group of a study and the intervention group. For example, if 30% of people experience a serious event in the control group and 20% in the intervention group, the ARR is 10% (30%-20%). Confidence intervals (CI): A statistical measure of the precision of estimation of a population (group) parameter at a specified probability level. These are usually 95% confidence intervals, meaning that 95% of times the value will lie within the range given, but there is a 5 in 100 chance that it will not, because of chance. Heterogeneity: A wide spread of results in a collection of data that suggests extra factors other than those that have already been taken into consideration are influencing the results. Number needed to treat (NNT): A statistical measure of the number of people who need to be given an intervention in order to observe a beneficial effect for one extra person. It is calculated from the risk difference. It is the inverse of the absolute risk difference. Odds: A way of expressing chance that is expressed by the number of events of interest divided by the number of observations or people without this event. It can be stated as ‘one person fell for every three that did not’ (1/3=0.33) and is best used when events are rare as it then gives a similar measure as risk. Odds ratio: A measure of the odds in the treatment group divided by the odds for a control group. A value of one implies no treatment effect. If the odds ratio is less than one then treatment has reduced the odds of an event (often expressed as a percent so that it gives a measure of how much in 100). Probability: A measure of the likelihood of occurrence of an event. P-value: A measure of the probability or likelihood that a given effect or event will take place by chance. The smaller the value the more likely that the intervention is responsible for an observed effect. Relative risk or risk ratio (RR): How often an outcome or event happens in the treatment group divided by how often it occurs in the control or second group. A value of one implies no effect of treatment; if less than one then the intervention reduced the risk of an event; if greater than one then the risk is increased. Risk: A way of expressing the chance of an event taking place, expressed as the number of events divided by the total number of observations or people. It can be stated as ‘the chances of falling were one in four’ (1/4 = 25%). Risk difference: How often an outcome or event is reported in the treatment group minus how often it happens in the control or other group. Standard deviation: A measure of the variability of a measurement and therefore the precision of a mean value. See also the Cochrane Collaboration glossary at http://www.cochrane.org/resources/glossary.htm Top of page Copyright © ccnet

Some of the language used in clinical trials and systematic reviews

Introduction

Looking into clinical trials, to participate or to look at the findings - here are some of the scientific language you are likely to come across.

Terminology

An interventional clinical trial includes all prospective research studies that assign humans or groups of humans to one or more interventions for the purpose of evaluating their effects on health outcomes.

Clinical trials are important sources of scientific evidence on the safety and effectiveness of health interventions. Access to information about ongoing, completed and published clinical trials is essential for appropriate decision-making. Researchers, research funders, policy-makers, medical practitioners, patients and the general public need such information, to help guide research or to make treatment decisions (www.who.int/ictrp).

Evidence-based health care is a combination of current best research evidence, the expertise of the healthcare provider, and patient values - making decisions about health care for and with individuals, families and communities.

The ‘gold standard’ for producing evidence on healthcare interventions is the double-blind, randomised clinical trial (RCT). An RCT randomises people to different treatments, as a way of trying to make the different groups as similar as possible. The intervention that is being tested is compared to a non-active (placebo), a known treatment or usual care. The people who give the treatment and assess the results of giving the treatments do not know who is receiving what intervention (double blinded). The larger the trial the more the influence of things we do not know about, including biological variations and behaviours from one person to another, can be reduced and the assessment of the effect of the treatment more accurate.

What an RCT does is measure the likely efficacy of a treatment – can the intervention be shown to work in this situation with carefully selected participants, a defined healthcare setting and careful follow up? The evidence from an RCT needs to be translated into everyday use in a social context and in busy healthcare systems. Furthermore, RCTs are usually short-term, so they tell us little about long-term benefits or harms of a healthcare intervention.

A systematic review of randomised controlled trials gathers individual trials and summarises them in a single review. A systematic review uses set methods to identify this best evidence and, where possible, pools the results in an overview or statistical meta-analysis. Studies are only included in the review if they meet specific, pre-specified criteria defined to minimise introducing biases and influencing the findings of both the individual studies and the review. Randomly assigning trial participants to an intervention group in a way that no-one is aware what group the person is in is an important part of these criteria. Many trials have insufficient numbers of participants to be able to distinguish intervention effects - because of natural variations between people and their responses to treatments (they are underpowered).

Some of the terms used in systematic reviews

Abstract: A brief, clearly structured summary of the main features and results of a study.

Adverse effect: An unfavorable, or even harmful, effect or event that occurs as a result of a healthcare intervention.

Allocation concealment: This is a process used in studies that involve different groups receiving different interventions or treatment. Ideally, allocation to the different groups is done in such a way that the participants, and the health care providers, do not know which intervention the participant is to receive. The intention is to avoid bias during the allocation process so that the intervention and control groups are as similar as possible.

Bias: Any factor, recognised or not, that distorts the findings of a study.

Blinding: A process where the participant, healthcare provider, or person assessing the outcomes (double blinding) are prevented from knowing which intervention the participant is receiving.

Characteristics of a study: Identifying features of a healthcare study that include the design of the study, participants, setting, interventions used, the outcomes monitored and follow-up of the people who were initially assigned to the study.

Confounder: A factor other than the intervention under investigation that could influence the study results and is sometimes difficult to identify. The factor is separately related to the outcome such that it distorts the relationship between intervention and outcome.

Controls or control group: The group in a study that does not receive the intervention that is being investigated. Controls may receive no intervention, a placebo, a commonly used treatment for that condition, or only a part of an intervention.

Controlled trial: An intervention trial in which a group given the intervention under study is compared to a control group which does not receive the intervention.

Effectiveness: The ability of an intervention to produce a desired effect in an individual.

Efficacy: The ability of an intervention to produce a desired effect in a controlled study.

Eligibility criteria: The criteria that are clearly set out beforehand to define whether a potential participant is eligible to enter a study. In the Cochrane context, this term refers to the criteria that are clearly set out beforehand to define whether a study meets the requirements for inclusion in a systematic review.

Epidemiology: The discipline which studies the causes, distribution, and contributing factors to healthcare problems in population or community groups.

Grey literature: Reports of studies that do not appear in the peer-reviewed literature. These reports are often not easily accessible and may not be in the format of papers in the general medical literature.

Intervention: A treatment, procedure or program of health care that has the potential to change the course of events of a healthcare condition.

Meta-analysis: A way of analysing information or data from a number of different studies to determine an average, or common, effect. It is aimed at improving the precision of the available data by looking at a greater number of people, including from different populations.

Morbidity: Illness or harm.

Objective(s): The clearly stated healthcare question(s) that a study or review sets out to answer.

Outcome: A measure of the effect of an intervention that is relevant to health care.

Participant(s): Individual(s) who take part in a study.

Placebo: An intervention that to all intents and purposes appears to be the same as that which is being assessed but which does not have the active component being assessed.

Placebo effect: A change due to an expectation that the treatment or procedure will have an effect rather than the treatment or procedure itself.

Population: This term can refer to the participants involved in a healthcare study; it can also refer to a general population of people.

Protocol: An outline of a healthcare question, its relevance, and how reviewers intend to go about answering the question. The process of developing a protocol is aimed at improving the quality of a future review and minimising bias by deciding beforehand the methods, selection criteria and outcomes to be assessed. A protocol is refereed (peer reviewed) and published in The Cochrane Library to invite feedback as to the usefulness of the question and the process as outlined.

Quasi-randomised controlled trial: This term is loosely used to refer to randomised controlled trials where the method of allocation to the different groups is not sufficiently rigorous to ensure allocation concealment. These methods include alternate patients and allocating patients by birth dates.

Randomisation: A method of allocation based on chance such that individual allocation cannot be anticipated and statistical theory can be used for analysis. It is a way of ensuring groups are as similar as possible at the beginning of a study.

Randomised controlled trial: A study design in which individuals are assigned, by special randomisation techniques, to two or more groups where one group receives the intervention under investigation and the other(s) receives no treatment, a placebo, or a standard intervention.

Search strategy: The way a search is carried out to identify publications or reports that are relevant to a review process. Bias can be minimised and the quality of the review improved by the thoroughness of the search.

Standard treatment: The treatment, procedure or intervention that is normally or conventionally given for a condition.

Statistical analysis: The process of examining collected numerical measurements in order to determine if they are a fair representation of a situation, what value or significance can be given to them, and consequently the size and numerical significance of an intervention effect.

Study: A systematic, well-planned investigation of a healthcare problem.

Subgroup: When participants of a study are further divided according to factors other than the intervention received eg age, sex, severity of disease or physical condition, dose of intervention or quality of study.

Some statistical terms

Absolute risk reduction (ARR) or risk difference: The difference in the rate of outcomes between the control group of a study and the intervention group. For example, if 30% of people experience a serious event in the control group and 20% in the intervention group, the ARR is 10% (30%-20%).

Confidence intervals (CI): A statistical measure of the precision of estimation of a population (group) parameter at a specified probability level. These are usually 95% confidence intervals, meaning that 95% of times the value will lie within the range given, but there is a 5 in 100 chance that it will not, because of chance.

Heterogeneity: A wide spread of results in a collection of data that suggests extra factors other than those that have already been taken into consideration are influencing the results.

Number needed to treat (NNT): A statistical measure of the number of people who need to be given an intervention in order to observe a beneficial effect for one extra person. It is calculated from the risk difference. It is the inverse of the absolute risk difference.

Odds: A way of expressing chance that is expressed by the number of events of interest divided by the number of observations or people without this event. It can be stated as ‘one person fell for every three that did not’ (1/3=0.33) and is best used when events are rare as it then gives a similar measure as risk.

Odds ratio: A measure of the odds in the treatment group divided by the odds for a control group. A value of one implies no treatment effect. If the odds ratio is less than one then treatment has reduced the odds of an event (often expressed as a percent so that it gives a measure of how much in 100).

Probability: A measure of the likelihood of occurrence of an event.

P-value: A measure of the probability or likelihood that a given effect or event will take place by chance. The smaller the value the more likely that the intervention is responsible for an observed effect.

Relative risk or risk ratio (RR): How often an outcome or event happens in the treatment group divided by how often it occurs in the control or second group. A value of one implies no effect of treatment; if less than one then the intervention reduced the risk of an event; if greater than one then the risk is increased.

Risk: A way of expressing the chance of an event taking place, expressed as the number of events divided by the total number of observations or people. It can be stated as ‘the chances of falling were one in four’ (1/4 = 25%).

Risk difference: How often an outcome or event is reported in the treatment group minus how often it happens in the control or other group.

Standard deviation: A measure of the variability of a measurement and therefore the precision of a mean value.

See also the Cochrane Collaboration glossary at http://www.cochrane.org/resources/glossary.htm

Top of page