Glossary

#- A- B- C- D- E- F- G- H- I- J- K- L- M- N- O- P- Q- R- S- T- U- V- W- X- Y- Z

Absolute Risk Reduction

Absolute Risk Reduction, commonly abbreviated to ARR, is the difference between the incidence of an outcome in the control group minus the incidence of an outcome in the treatment group (Centre for Evidence Based Medicine, 2005). For example, if the number of events (e.g. death) in the control group (e.g. receiving placebo) is 40 per 100 patients, while the number of events in the treatment/intervention group is 30 per 100 patients, ARR = 0.40 – 0.30; ARR = 0.10 or 10% (Barratt et al., 2004). This means the absolute benefit of treatment is a 10% reduction in the death rate.

|An ARR of 0 means that there is no treatment effect.|

allocation concealment

A procedure used to ensure that the person entering participants into a study does not know the participants will be in the intervention group or in the control group. Allocation concealment differs from blinding, in that it aims to prevent selection bias, rather than performance bias (The Cochrane Collaboration, 2010). In addition, one can always implement allocation concealment, whereas blinding is not always possible.

ARR

ARR, the common abbreviation of Absolute Risk Reduction, is the difference between the incidence of an outcome in the control group minus the incidence of an outcome in the treatment group (Centre for Evidence Based Medicine, 2005). For example, if the number of events (e.g. death) in the control group (e.g. receiving placebo) is 40 per 100 patients, while the number of events in the treatment/intervention group is 30 per 100 patients, ARR = 0.40 – 0.30; ARR = 0.10 or 10% (Barratt et al., 2004). This means the absolute benefit of treatment is a 10% reduction in the death rate.

|An ARR of 0 means that there is no treatment effect.|

bias

Any procedural error or influence, occurring at any point in a study, that leads to conclusions that differ from the truth (Gay, 2010). Typical sources of bias are:

Differences between the groups under comparison (selection bias)
Inconsistencies with the care/intervention that is provided
Exposure to other factors apart from the intervention of interest (performance bias)
Withdrawals or exclusions of people entered into a study (attrition bias)
How outcomes are assessed (detection bias)

Reviews of primary studies may also be affected by reporting bias, where a only a portion of all the relevant data is made available (The Cochrane Collaboration, 2010).

BIOSIS

A bibliographic database produced by Thomson Scientific, that indexes literature related to biological and biomedical sciences.

blinding

Blinding, in an experimental study, refers to whether patients, clinicians providing an intervention, people assessing outcomes, and/or data analysts knew which groups patients were assigned to (BMJ Publishing Group Ltd. & RCN Publishing Company Ltd., 2009) An example of a double-blind study would be one in which the researchers randomly assign study participants (i.e. patients) an active drug or placebo, and neither the patient nor his/her attending clinician is informed of which the patient received. Theoretically, this allows the clinician to assess patient outcomes more objectively. Blinding decreases the degree of bias, though it is not always appropriate or feasible in some studies.

case-control study

A research design that compares individuals with an outcome of interest (e.g. heart disease) to individuals from the same population without that outcome by retrospectively gathering information about study subjects’ prior exposure to the factor(s) under examination (e.g. environmental tobacco smoke) (The Cochrane Collaboration, 2010; Law & Howick, 2011). Case-control studies are a type of retrospective, observational study, and typically generate odds ratios.

CI, the common abbreviation of confidence interval, provides the likely range of the true value of interest (e.g. the effect of an intervention or treatment). The CI is usually reported as '95% CI', which means that 95% of the time the true value for the population lies within the given range of values (BMJ Publishing Group Ltd. & RCN Publishing Company Ltd., 2009). A CI of 90% or 99% may also be used. A wider CI means the sample is more variable and the point estimate (i.e. estimate of true value; see point estimate) is less precise.

The results of a study are statistically significant when its confidence interval does not include the value corresponding to no difference between values of interest (e.g. the null effect: 1 for odds and risk ratios , or 0 for mean differences) (DiCenso, 2001). Take, for example, a relative risk reported as 1.03 with a CI (0.85 - 1.26). The CI captures the value of 'no effect' (i.e. 1.0), as it runs from 0.85 to 1.26. Therefore, the observed difference is not statistically significant; the true value could be anything from a 15% reduction in events (e.g. death), with a given intervention, to a 26% increase in events (Davies & Crombie, 2009).

|NOTE: There is only one true value. The confidence interval defines the range where it's most likely to be.|

CINAHL

A bibliographic database that indexes literature related to nursing and midwifery as well as literature from the allied health disciplines such as physiotherapy, health education and nutrition.

clinical significance

A measurement of how effective a treatment/intervention is when applied to real-world scenarios. Assessing clinical significance takes into account factors such as the size of a treatment effect, the severity of the condition being treated, the side effects of the treatment, and the cost (Leung, 2000).

Cochran chi-square (Cochran’s Q) test

Cochran’s Q, distributed as a chi-square (X²) statistic, examines the null hypothesis that all studies of a review are examining the same effect; in other words, whether observed differences in results are the result of chance alone (Higgins & Green, 2009; Higgins, Thompson, Deeks, & Altman, 2003). It is calculated as the weighted sum of squared differences between individual study effects and the pooled effect across studies. Afterwards, a p-value is calculated (see p-value). Cochran’s Q is not considered the most effective test of heterogeneity. It has poor power (i.e. cannot detect heterogeneity that is clinically important) when there are only a few studies and excessive power (i.e. begins to detect heterogeneity that is clinically unimportant) when there are many studies (Higgins & Green, 2009).

|A p-value < 0.05 (or sometimes < 0.10) means heterogeneity exists between studies (i.e. reject the null hypothesis) (Higgins et al., 2003). A p-value of 0.01, for example, means that there is only a 1% chance that this deviation is due to chance alone.|

|A p-value > 0.05 (or sometimes > 0.10) means there is no heterogeneity and chance alone accounts for differences between studies (i.e. accept the null hypothesis) (Higgins et al., 2003). A p-value of 0.6, for example, means that there is a 60% probability that any deviation is due to chance only.|

cohort

A group of individuals who are linked in some way or who have experienced the same significant life event within a given period (e.g. birth year, or exposure to a particular drug).

cohort study

A research design that compares two or more groups, in which at least one group has been exposed to a risk factor/intervention/treatment (e.g. chemical fertilizer) and one has not (Gay, 2010; Law & Howick, 2011). Typically the groups are followed forward in time (prospective) to track the incidence of different outcomes (e.g. cancer). In a retrospective cohort study, cohorts are identified by the researcher(s) based on their exposure to a risk factor/intervention/treatment at some point in the past and information is collected, often from medical records, on subsequent outcomes (Buckingham, Fisher, & Saunders, 2008). Cohort studies are a type of observational study, and are ideal for examining the effect of predictive risk factors on an outcome (Law & Howick, 2011).

Confidence Interval

continuous data

Data which are numerical and made up of many ordered categories (The Cochrane Collaboration, 2010). Examples include body weight and blood pressure.

controlled before-and-after study

A non-randomised study design where a control population of similar characteristics and performance as the intervention group is identified. Data are collected before and after the intervention in both the control and intervention groups (The Cochrane Collaboration, 2010; SUPPORT, n.d.). Before-and-after studies can also be uncontrolled, and would be at higher risk of bias.

degrees of freedom (d.f.)

The number of independent contributions (e.g. subjects, test scores, conditions) that can be made to a data sample and a measure of how certain we are that our sample population is representative of the entire population. Therefore, the more degrees of freedom, the more representative the sample (The Cochrane Collaboration, 2010). The formula used to calculate degrees of freedom depends on the statistical test (e.g. F test) that one is using.

diagnostic study

A research design that assesses the ability of a test or measurement (e.g. mammography) to detect whether someone has (or does not have) a specific disease (National Insitute for Health and Clinical Excellence, 2010). In some cases, a diagnostic study tests whether a new diagnostic method is as effective as the current ‘gold standard’ method for diagnosing a given disease (National Health Service, 2009).

dichotomous data

Data for which an outcome is one of only two possible categorical responses. Examples include, “deceased or alive,” “male or female,” etc. (Higgins & Green, 2009)

dissemination

The one-way spreading of knowledge or research, such as is done in scientific journals and at scientific conferences as well as extracting the main messages or key implications derived from research results and communicating them to a specific audience (Calgary Health Region, 2006; The Canadian Health Services Research Foundation, n.d.)

effect size

Effect size is a generic term for the estimate of the true value of the effect (i.e. the amount of change) from a given intervention compared to not receiving the treatment/intervention or receiving another intervention (DiCenso, 2001; Higgins & Green, 2009). Effect size may be expressed as a relative risk, odds ratio, relative risk reduction, etc.

effectiveness

The measure of the ability of an intervention, project, program, or policy to do what it was intended to do: produce a specific desired result or effect that can be quantitatively measured (European Observatory on Health Systems and Policies, n.d.).

EMBASE

A bibliographic database that indexes drug and biomedical literature.

ERIC

ERIC (Education Resources Information Center) is an online digital library of education research and information.

evaluate

Efforts aimed at determining, as systematically and objectively as possible, the relevance, effectiveness and impact of health-related (and other) activities in relation to objectives, taking into account the resources that have been used (Public Health Agency of Canada, 2006; Rychetnik, Hawe, Waters, Barrat, & Frommer, 2004).

evidence

Knowledge from a variety of sources including qualitative and quantitative research, program evaluations, client values and preferences, and professional experience (Calgary Health Region, 2006).

evidence summary

A concise, plain-language summary of a review that describes the review's main topic area (framed in a Canadian context), content, methods and findings, and provides implications for public health practice, programs, and policy.

evidence-informed decision making

The purposeful and systematic use of the best available evidence to inform the assessment of various options and related decision making in practice, program development, and policy making. This process involves searching for, accessing, assessing the relevance and quality of evidence; interpreting this evidence and identifying associated implications for practice, program and policy decisions; adapting this evidence in light of the local context; implementing this evidence; and evaluating its impact (Calgary Health Region, 2006; Canadian Health Services Research Foundation, n.d.; DiCenso, Guyatt, & Ciliska, 2005; Sackett, Richarson, Rosenberg, & Haynes, 1997).

experimental study

A research design, most often a randomized controlled trial, that sorts study participants into two or more groups, with a minimum of one control group, the other(s) being the intervention group [BMJ Publishing Group, 2011; National Institute for Health and Clinical (NICE) Excellence, 2010]. Groups are then followed to assess whether or not an intervention (e.g. test, or treatment) that is controlled by the investigator, affects the course or outcome of a condition or disease (NICE, 2010). Experimental studies (rather than observational studies) are the preferred choice for most medical studies because control groups help minimize bias (NICE, 2010).

F-statistic

A value, resulting from the F-test and regression analysis, used to determine if the differences between the means of two populations are different enough to be statistically significant. The F-statistic is not directly used for clinical interpretation (Barlow Pugh, 2006).

guideline

Public health guidelines are recommendations for populations and individuals on activities, policies and strategies that can help prevent disease or improve health [National Institute for Health and Clinical Excellence (NICE), 2010]. The guidance may focus on a particular topic (e.g. cancer), a particular population (e.g. seniors) or a particular setting (e.g. home). They are based on the best available evidence. While guidelines help health professionals in their work, they do not replace their knowledge and skills (NICE, 2010).

hazard ratio

This represents the increased risk with which one group is likely to experience the outcome of interest. For example, if the hazard ratio for death for a treatment is 0.5, then we can say that treated patients are likely to die at half the rate of untreated patients (The Cochrane Collaboration, 2010). A hazard ratio of 2.0 means that treatment will cause the patient to progress more quickly, and that a treated patient who has not yet progressed by a certain time has twice the chance of having progressed at the next point in time compared with someone in the control group.

health promotion

The comprehensive social and political process of enabling people to increase control over and improve their health through actions aimed at strengthening individual awareness and skill; changing individual behaviour; and changing social, organizational, political, and economic conditions that support good health practices (O' Donnell, 1989; Stanhope & Lancaster, 1996; World Health Organization, 1998).

heterogeneity tests

Heterogeneity tests measure the extent of differences (i.e. heterogeneity) between individual studies. A significant test of heterogeneity suggests that the observed differences between individual study results are not the result of chance alone: another factor (e.g. study design, population) is responsible for the differences in treatment/intervention effects across studies (National Institute for Health and Clincal Excellence, 2010). The results of these tests often help review authors decide whether it is appropriate to perform statistical synthesis (i.e. conduct a meta-analysis). Examples of heterogeneity tests include Cochran's Q, and the I² test of heterogeneity

I²

The I² test of heterogeneity describes the percentage of total variation across studies due to heterogeneity rather than chance (Higgins, Thompson, Deeks, & Altman, 2003). An I² value of 0% indicates no heterogeneity. In general, an I² value of 25% reflects low heterogeneity, 50% moderate heterogeneity, and 75% high heterogeneity (Higgins, Thompson, Deeks, & Altman, 2003). The I² can be calculated from a variety of study designs and types of outcome data. It is effective at detecting true heterogeneity between a small number of studies (Higgins, Thompson, Deeks, & Altman, 2003). I² = 100%x(Q – df)/Q (where Q is Cochran’s Q and df the degrees of freedom).

incidence

The number of new occurrences of something in the total population during a certain period. For example, the number of new cases of a disease in a country over one year (The Cochrane Collaboration, 2010; Law & Howick, 2011; National Institute for Health and Clinical Excellence, 2010). (See prevalence)

incidence rate

Refers to the number of new cases of a condition in a particular group or population. For example, if there are 1000 people and 14 of them develop a condition, the incidence rate is 14 per 1000 or 1.4% (Harvard School of Public Health, 2004). (See incidence.)

intention-to-treat analysis

A strategy for analyzing data from a randomized controlled trial in which all participants are assessed based on the comparison group to which they were originally allocated, whether or not they received (or fully complied with) the intervention given to that group (BMJ Publishing Group, 2011; The Cochrane Collaboration, 2010; National Institute for Health and Clinical Excellence, 2010). Intention-to-treat analysis helps prevent bias caused by the loss of participants, and the resulting imbalance between comparison groups.

interrupted time series

A research design that collects data at multiple time points before and after an intervention or natural event (i.e. interruption) (The Cochrane Collaboration, 2010).

intervention

The aspect of interest in experimental and observational studies. Interventions can be therapeutic (e.g. different wound dressings), preventative (e.g. influenza vaccination), and diagnostic (e.g. measurement of blood pressure), targeted at individuals, groups, organizations, communities or health systems.

IRR

The common abbreviation of Incidence Rate Ratio, IRR is the ratio of two incidence rates. The incidence rate among the exposed - to a risk factor/treatment/intervention - proportion of a particular population, divided by the incidence rate in the unexposed portion of this population, gives a relative measure of the effect of a given exposure (Harvard School of Public Health, 2004).

knowledge broker

An individual or organization that aims to develop relationships and networks with, among, and between producers and users of evidence and knowledge in order to facilitate: a) knowledge exchange and co-development, b) the appropriate use of the best available evidence in decision-making processes, and c) individual and organizational capacity to participate effectively in this evidence-informed decision making process.

knowledge management

“The purpose of knowledge management is to provide support for improved decision making and innovation throughout the organization. This is achieved through the effective management of human intuition and experience augmented by the provision of information, processes and technology together with training and mentoring programmes” (Snowden, 2009).

knowledge transfer and exchange

A two-way process involving dialogue, interaction, and the sharing of knowledge and evidence between and among producers and users of that evidence or knowledge resulting in mutual learning through the process of planning, producing, disseminating, and applying existing or new research in decision-making and evaluating the impacts of the related decisions. Abbreviated as KTE, this is a broad term that is often used to include knowledge transfer, exchange, translation, dissemination, and diffusion (Canadian Health Services Research Foundation, n.d.).

mean

A value calculated by adding each value in a set of numbers/observations and dividing by the total number of observations (The Cochrane Collaboration, 2010). The mean is commonly called the ‘average.’

|For example, given the set of numbers 1, 9, 8, 6, 3 the mean = 5.4 [(1+3+6+8+9)/5]|

mean difference

The mean difference (or 'difference in means') is a statistic measuring the absolute difference between the mean value in two groups in a clinical trial (Higgins & Green, 2009). It estimates the amount by which the experimental intervention changes the outcome on average, compared with the control. It can be used as a summary statistic in meta analysis when continuous outcomes (e.g. height) are measured on the same scale in ALL included studies (BMJ Publishing Group Ltd., & RCN Publishing Company Ltd., 2009; Higgins & Green, 2009). (See also Standardized Mean Difference.)

median

The observation/value that comes half way when the observations are ranked in sequential order (The Cochrane Collaboration, 2010). This will divide the set of observations/values into an equal quantity below and above the median value.

|For example, given the set of numbers 1, 3, 6, 8, 9 the median = 6|

MEDLINE

A bibliographic database that indexes biomedical literature.

meta-analyses

A statistical technique to combine the results of multiple studies resulting in a single pooled estimate of effect.

moderate

All reviews in Health Evidence are assessed for methodological quality. Reviews receive a quality assessment score out of 10. Reviews with a score between 5-7 are rated as moderate.

narrative review

Narrative reviews are evidence overviews or expert commentaries on a given health topic. Unlike systematic reviews, narrative reviews are not designed to be reproducible as their methodology (e.g. search strategy, inclusion criteria) is usually not described. As such, narrative reviews are vulnerable to bias.

NNH

NNH, the common abbreviation of Number Needed to Harm, is the number of patients who must receive a treatment or intervention to cause one adverse event [DiCenso, 2001; National Institute for Health and Clinical Excellence (NICE), 2010]. For example, if you give a stroke prevention intervention to 100 people and 4 of them experience joint pain (i.e. an adverse event), the number needed to harm is 25 (that is, 100 divided by 4 equals 25) (NICE, 2010).

|The closer the NNH is to 1, the more likely it is that someone on the treatment will experience an adverse event (NICE, 2010)|

NNT

NNT, the common abbreviation of Numbers Needed to Treat, is the number of patients who must be treated to prevent 1 additional negative event (or to promote 1 additional positive event). This is calculated as 1/absolute risk reduction (rounded to the next whole number), accompanied by the 95% confidence interval (BMJ Publishing Group Ltd., & RCN Publishing Company Ltd., 2009; Buckingham, Fisher, & Saunders, 2008). When assessing clinical significance, one must consider the NNT against any harms or adverse effects. (See Numbers Needed to Harm.)

|From the ARR example NNT = 1/0.10 = 10 Therefore, one expects that 10 people require the experimental intervention in order to prevent one event (e.g. death) (Higgins & Green, 2009).|

NOTE: The NNT gives an ‘expected value’. For example, NNT = 10 does not imply that one additional event will occur in each and every group of ten people (Higgins & Green, 2009).

null hypothesis

The null hypothesis is the statistical hypothesis that one variable (e.g. the treatment or intervention) has NO association with another variable or set of variables (e.g. an outcome, such as death), or that two or more population distributions do NOT differ from one another (Higgins & Green, 2009). For example, a particular vaccine may be used to prevent influenza. A possible null hypothesis may be "this vaccine has no effect on the incidence of influenza."

Number Needed to Harm

Number Needed to Harm, commonly abbreviated to NNH, is the number of patients who must receive a treatment or intervention to cause one adverse event [DiCenso, 2001; National Institute for Health and Clinical Excellence (NICE), 2010]. For example, if you give a stroke prevention intervention to 100 people and 4 of them experience joint pain (i.e. an adverse event), the number needed to harm is 25 (that is, 100 divided by 4 equals 25) (NICE, 2010).

|The closer the NNH is to 1, the more likely it is that someone on the treatment will experience an adverse event (NICE, 2010).|

Number Needed to Treat

Number Needed to Treat, commonly abbreviated to NNT, is the number of patients who must be treated to prevent 1 additional negative event (or to promote 1 additional positive event). This is calculated as 1/absolute risk reduction (rounded to the next whole number), accompanied by the 95% confidence interval (BMJ Publishing Group Ltd., & RCN Publishing Company Ltd., 2009; Buckingham, Fisher, & Saunders, 2008). When assessing clinical significance, one must consider the NNT against any harms or adverse effects. (See Numbers Needed to Harm.)

|From the ARR example NNT = 1/0.10 = 10 Therefore, one expects that 10 people require the experimental intervention in order to prevent one event (e.g. death) (Higgins & Green, 2010).|

NOTE: The NNT gives an 'expected value’. For example, NNT = 10 does not imply that one additional event will occur in each and every group of ten people (Higgins & Green, 2009).

observational study

A research design that requires investigators do not intervene or control variables rather, they simply observe the course of events. Changes or differences in one characteristic (e.g. whether or not people received the intervention of interest) are studied in relation to changes or differences in other characteristic(s) (e.g. death), without action by the investigator. There is a greater risk of bias in observational studies than in experimental studies (The Cochrane Collaboration, 2010).

Odds Ratio

Odds Ratio, commonly abbreviated to OR, is calculated by dividing the odds of an event occurring in the intervention group by the odds of that event occurring in the control group or the odds that a patient was exposed to a given risk factor divided by the odds that a control patient was exposed to the risk factor (BMJ Publishing Group Ltd., & RCN Publishing Company Ltd., 2009; Moher, Jadad, & Klassen, 1998).

For example, if the number of events (e.g. death) in the control group (e.g. receiving placebo) is 60 per 100 control group members, the odds are 60 to 40 (i.e. 60 people died, 40 people did not). If the number of events in the treatment or intervention group is 15 per 100 intervention group members, the odds are 15 to 85. This means that: OR = (15/85) / (60/40) = 0.12. Thus, the intervention reduced the odds of death by 12%, or the intervention reduced the odds of death by 88% of what they were in the control group (The Cochrane Collaboration, 2010).

On the other hand, benefit of an interventions given to encourage favourable outcomes (e.g. pregnancy in infertile couples) is reflected by an OR > 1.0

NOTE: Unlike a Relative Risk, an Odds Ratio does NOT tell us the probability of an event happening. When an outcomes is rare, however, the Odds Ratio is often similar to the Relative Risk.

|An OR value of 1 means that there is NO DIFFERENCE between the intervention and control groups|

The common abbreviation of Odds Ratio, OR is calculated by dividing the odds of an event occurring in the intervention group by the odds of that event occurring in the control group or the odds that a patient was exposed to a given risk factor divided by the odds that a control patient was exposed to the risk factor (BMJ Publishing Group Ltd., & RCN Publishing Company Ltd., 2009; Moher, Jadad, & Klassen, 1998).

For example, if the number of events (e.g. death) in the control group (e.g. receiving placebo) is 60 per 100 control group members, the odds are 60 to 40 (i.e. 60 people died, 40 people did not). If the number of events in the treatment or intervention group is 15 per 100 intervention group members, the odds are 15 to 85. This means that: OR = (15/85) / (60/40) = 0.12. Thus, the intervention reduced the odds of death by 12%, or the intervention reduced the odds of death by 88% of what they were in the control group (The Cochrane Collaboration, 2010).

On the other hand, benefit of an interventions given to encourage favourable outcomes (e.g. pregnancy in infertile couples) is reflected by an OR > 1.0

NOTE: Unlike a Relative Risk, an Odds Ratio does NOT tell us the probability of an event happening. When an outcomes is rare, however, the Odds Ratio is often similar to the Relative Risk.

|An OR value of 1 means that there is NO DIFFERENCE between the intervention and control groups|

ordinal data

Data that are classified into more than two categories which have an intuitive order; for example, pain scale of “none/mild/moderate/severe” (The Cochrane Collaboration, 2010). Generally, short scales (e.g. a 1-5 pain scale) are analyzed as ordinal data.

PICO

An acronym for: Population, Intervention (e.g. exposure, screening tool, therapy), Comparison/control, and Outcome (BMJ Publishing Group, 2011). PICO statements are often used to generate a quantitative clinical question or arise from a practice-based scenario or issue, and then used to guide a review of the health literature.

For example:

P: Who is the population/patient/problem?
--- Healthy children

I: What is the intervention/exposure/risk/prognostic factor of interest?
--- Annual flu shot

C: Is there a comparison or control group?
--- Healthy children not receiving the annual flu shot

O: What is the outcome(s) of interest?
--- Decreased rates of influenza in healthy children

The clinical question for the above PICO may be: For healthy children, how effective is the annual flu vaccination in decreasing influenza rates when compared to no vaccination?

point estimate

The point estimate is the statistical best guess or estimate of the effect of an intervention or treatment (e.g. the true value of interest).

population health

Population health is an approach to health that aims to improve the health of the entire population and to reduce health inequities among population groups. In order to reach these objectives, it looks at and acts upon the broad range of factors and conditions that have a strong influence on our health (Public Health Agency of Canada, 2002).

power

Power is the ability of a test to reject the null hypothesis when a specific alternative hypothesis is true (The Cochrane Collaboration, 2010). In clinical trials, power is the probability that a trial will detect, as statistically significant, an intervention effect. If a clinical trial reports a power of 0.80 (or 80%), and assuming that the pre-specified treatment effect truly existed, then if the trial was repeated 100 times, one would find a statistically significant treatment effect in 80 of them (The Cochrane Collaboration, 2010). Ideally we want a test to have high power, close to maximum of one (or 100%). In general, studies with more participants have greater power (The Cochrane Collaboration, 2010).

prevalence

The proportion of a population having a particular condition or characteristic. For example, the percentage of people in a city with a particular disease, or who smoke (The Cochrane Collaboration, 2010; National Institute for Health and Clinical Excellence, 2010). (See incidence.)

An acronym for: Patient/population, and Situation (National Collaborating Centre for Methods and Tools, n.d.). PS statements are often used to generate a qualitative research question or arise from a practice-based scenario or issue, and then used to guide a review of the health literature.

For example:

P: How would you describe the patient/population of interest? Do they have important characteristics? What are their age, gender, and/or socioeconomic status?
--- First-time mothers with an infant < 6 months of age

S: What circumstances or situations do you want to know about?
--- What influenced their decision to breastfeed or use infant formula?

The clinical question for the above PS may be: Why do first-time mothers choose to breastfeed or infant formula?

PsycINFO

A bibliographic database that indexes literature related to psychology and related disciplines.

public health

A combination of sciences, skills, and values that function through collective societal, legislative, and political activities and involve programs, services, and institutions aimed at protecting and improving the health of all people a scientific, technical, social, and political endeavour, involving all organized measures both public and private that aim to prevent disease, promote health and wellbeing, prolong life, and when necessary, restore the health of individuals, specified groups, populations or communities through mobilizing and engaging local, state, national, and international resources to assure the conditions in which people can be healthy (Public Health Agency, 2006; Rychetnik, Hawe, Waters, Barratt, & Frommer, 2004; World Health Organization, 2006).

public health decision makers

Decision makers in the health services field can range from frontline health providers to administrators to ministers of health. Health Evidence focuses its efforts on two types of decision makers: managers and policy makers who often work in public health organizations such as health units and regional health authorities, non-governmental organizations, as well as ministries of health and relevant regulatory agencies (Canadian Health Services Research Foundation, n.d.).

public health interventions

An intervention applied to many, most, or all members in a community, with the aim of delivering a net benefit to the community or population as well as benefits to individuals (Rychetnik, Hawe, Waters, Barratt, & Frommer, 2004).

p-value

p-values represent the possibility that any particular outcome would have occurred by chance. Confidence intervals, however, are considered a better indicator of significance (Buckingham, Fisher, & Saunders, 2008). The p-value gives the probability of obtaining the present test result—or an even more extreme one—if there were no effect from a treatment or intervention. A small p-value (i.e. p<0.05) signifies that the probability is small that the finding from the study occurred due to chance; rather, this finding would only be observed 5 times out of 100 and be incorrect.

|Statistical significance is usually set at p<0.05, although p<0.01 or p<0.001 may be used in instances where the outcome is serious (e.g. death).|

qualitative

Qualitative research is used to explore and understand people's beliefs, experiences, attitudes, behaviour and interactions. It generates non- numerical data (e.g., a patient's description of their pain rather than a measure of pain). In health care, qualitative techniques have been commonly used in research documenting the experience of chronic illness and in studies about the functioning of organizations (Bandolier, n.d.).

quality assessment

Quality assessment criteria are checklists or standards that are used to evaluate research evidence. These criteria can be applied to assess the value of a single study, or they are used to assess several studies as part of the process of systematic review. Quality assessment criteria address different variables, depending on the nature and purpose of the research, and the expectations and priorities of the reviewers. Items included are: methodological rigour, levels of evidence, strength of evidence, magnitude, completeness, relevance, and criteria of causation (Rychetnik, Hawe, Waters, Barratt, & Frommer, 2004).

quasi-experimental study

A research design that differs from experimental studies in that participants are not randomly assigned to groups, but the investigator still controls the intervention(s) (e.g. test, or treatment) received by at least one of the groups. This means a researcher can't draw conclusions about 'cause and effect'. This design is frequently used when it is not feasible, or not ethical, to conduct a randomized controlled trial (National Institute for Health and Clinical Excellence, 2010).

random allocation

Also known as random sampling or randomization, random allocation is a method that uses chance to allocate participants to comparison groups in a study. Researchers may use a computer-generated random sequence, for example. The randomization process implies that each individual/unit entering a given study has the same chance of receiving each of the possible interventions (The Cochrane Collaboration, 2010). This process also implies that the probability that an individual will receive a particular intervention is independent of the probability that any other individual will receive the same intervention (The Cochrane Collaboration, 2010).

randomized controlled trial

An experiment in which participants or populations are allocated by chance to receive an intervention (intervention group) or not (comparison or control group) and then followed up over time to determine the effect of that intervention by assessing differences in outcome rates (DiCenson, Guyatt, & Ciliska, 2005; Waters et al., 2006).

regression analysis

A statistical technique used to estimate or predict the influence of one or more independent variable (e.g. sex, age, education level) on a dependent variable (e.g. prevalence of a disease). Logistic regression and meta-regression are types of regression analysis (The Cochrane Collaboration, 2010).

Relative Risk

Relative risk, commonly abbreviated to RR, is the proportion of patients experiencing an outcome in the treated (or exposed) group divided by the proportion experiencing the outcome in the control (or unexposed) group (BMJ Publishing Group Ltd., & RCN Publishing Company Ltd., 2009). It tells us the risk or probability of an event occurring in the intervention group compared to the control group.

|A RR of 1 means there is NO SIGNIFICANT DIFFERENCE between the intervention and control groups (BMJ Publishing Group Ltd., & RCN Publishing Company Ltd., 2009)
For example, if the number of events (e.g. death) in the control group (e.g. receiving placebo) is 60 per 100 control group members, while the number of events in the treatment/intervention group is 15 per 100 intervention group members, RR = (15/100) / (60/100), RR = 0.25 or 25%. Treatment reduced the probability(risk) of an event to 25% of what it would have been without treatment (Higgins & Green, 2009). An alternative interpretation is the Relative Risk Reduction|

|A RR < 1.0 means the treatment decreases the risk of the outcome occurring.|

|A RR > 1.0 means that the treatment increases the likelihood of the outcome occurring.|

Relative Risk Reduction

Relative Risk Reduction, commonly abbreviated to RRR, is reported as a percentage and represents the extent to which a treatment or intervention reduces a person's risk of experiencing a given outcome (BMJ Publishing Group Ltd., & RCN Publishing Company Ltd., 2009; Buckingham, Fisher, & Saunders, 2008). It is calculated by subtracting the relative risk from 1 (i.e. the risk of the outcome), (RRR = 1 - RR) with accompanying confidence intervals. From the Relative Risk example above, RRR = (1 - 0.25), RRR = 0.75 or 75%. In this example, the treatment or intervention reduced the risk of death by 75% relative to that occurring in the control group. RRR is preferred over Absolute Risk Reduction, particularly when the event rate in the control group is low (Barratt et al., 2004).

|A RRR of 100% means that the treatment was a complete success|

|A RRR of 0% means that the treatment had no effect|

The common abbrevation of Relative Risk, RR is the proportion of patients experiencing an outcome in the treated (or exposed) group divided by the proportion experiencing the outcome in the control (or unexposed) group (BMJ Publishing Group Ltd., & RCN Publishing Company Ltd., 2009). It tells us the risk or probability of an event occurring in the intervention group compared to the control group.

|A RR of 1 means there is NO SIGNIFICANT DIFFERENCE between the intervention and control groups (Centre for Evidence Based Medicine, 2005). |

For example, if the number of events (e.g. death) in the control group (e.g. receiving placebo) is 60 per 100 control group members, while the number of events in the treatment/intervention group is 15 per 100 intervention group members, RR = (15/100) / (60/100), RR = 0.25 or 25%. Treatment reduced the probability(risk) of an event to 25% of what it would have been without treatment (Higgins & Green, 2009). An alternative interpretation is the Relative Risk Reduction

|A RR < 1.0 means the treatment decreases the risk of the outcome occurring.|

|A RR > 1.0 means that the treatment increases the likelihood of the outcome occurring.|

RRR

RRR, the common abbreviation of Relative Risk Reduction, is reported as a percentage and represents the extent to which a treatment or intervention reduces a person's risk of experiencing a given outcome (BMJ Publishing Group Ltd., & RCN Publishing Company Ltd., 2009; Buckingham, Fisher, & Saunders, 2008). It is calculated by subtracting the relative risk from 1 (i.e. the risk of the outcome), (RRR = 1 - RR) with accompanying confidence intervals. From the Relative Risk example, RRR = (1 - 0.25), RRR = 0.75 or 75%. In this example, the treatment or intervention reduced the risk of death by 75% relative to that occurring in the control group. RRR is preferred over Absolute Risk Reduction, particularly when the event rate in the control group is low (Barratt et al., 2008).

|A RRR of 100% means that the treatment was a complete success|

|A RRR of 0% means that the treatment had no effect|

The acronym for standard deviation, SD is a measure of how far individual data points are from the mean value for the set of data points (Gay, 2010; SUPPORT, n.d.). In other words, measures the spread of data across a sample, so a large standard deviation suggests data are spread out over a wide range of values.

|The average (i.e. mean) height of males in a given population is reported as 5'10" ± 3". In this case the SD is 3, so the range of height within 1 SD of the mean is 5’7” – 6’|

In any normal distribution, roughly two-thirds (actually, 68.2%) of the scores fall between -1 and +1 SD, and 95.4% between -2 and+2 SD

Sociological Abstracts

A bibliographic database that indexes the international literature in sociology and related disciplines in the social and behavioural sciences.

Sport Discus

A bibliographic database that indexes literature related to sport, health, fitness and sports medicine.

stakeholders

All persons, agencies and organizations with a direct or indirect interest in the development of an intervention or its evaluation. Those with an investment or 'stake' in the health of the community and the local public health system. This broad definition includes persons and organizations that benefit from and/or participate in the delivery of services that promote the public's health and overall well-being (Mobilizing for Action through Planning and Partnership, n.d.).

standard deviation

A measure of how far individual data points are from the mean value for the set of data points (Gay, 2010; SUPPORT, n.d.). In other words, standard deviation (SD), measures the spread of data across a sample, so a large standard deviation suggests data are spread out over a wide range of values.

|The average (i.e. mean) height of males in a given population is reported as 5'10" ± 3". In this case the SD is 3, so the range of height within 1 SD of the mean is 5’7” – 6’|

In any normal distribution, roughly two-thirds (actually, 68.2%) of the scores fall between -1 and +1 SD, and 95.4% between -2 and+2 SD

standard error of mean

The standard error of mean (SEM) is the precision of the estimate of a sample mean. SEM is a measure of the variability or spread of the means from repeated samples (of a given size) drawn from the population (Gay, 2010). In other words, SEM, measures the range of sample means.

Mathematically, the SEM is the standard deviation (SD) divided by the square root of the sample size, so it is always smaller than the SD. SEM cannot be used in place of SD (Gay, 2010).

Standardized Mean Difference

Standardized Mean Difference, commonly abbreviated to SMD, appears in the meta-analyses of continuous data (e.g. weight) when studies use different scales of measurement for the same outcome (e.g. weight in kilograms vs. weight in pounds) (The Cochrane Collaboration, 2010). It is necessary to standardize individual study results on a uniform scale so they can be combined (The Cochrane Collaboration, 2010). SMDs do not have units. Hedges’ (adjusted) g is the preferred SMD for Cochrane reviews. To calculate, SMD = difference in mean outcome between groups/standard deviation of outcome among participants.

For example, consider a trial evaluating an intervention to increase birth weight. The mean birth weights in intervention and control groups were 2700g and 2600g, respectively, with an average SD of 500g. SMD = (2700 - 2600)/500 = 0.2

statistical significance

An effect size is statistically significant when any differences in outcome(s) between treatment and control groups are likely real, and not due to chance. p-values and confidence intervals (CI) are the most commonly used measures of statistical significance (Leung, 2000).

stratification

The process of separating study population/sample into subsamples according to an important characteristic (e.g. gender, weight, or age) (Barlow Pugh et al., 2006).

strong

All reviews in Health Evidence are assessed for methodological quality. Reviews receive a quality assessment score out of 10. Reviews with a score of 8 or higher are rated as strong.

surveillance

In public health, surveillance, refers to the ongoing and systematic collection, analysis and interpretation of health-related data required for the planning, implementation, and evaluation of public health policy and practice (World Health Organization, 2011).

systematic review

A review of a clearly formulated question that uses systematic and explicit methods to identify, select, and critically appraise relevant research, and to collect and analyze data from the studies that are included in the review. Statistical methods (meta-analysis) may or may not be used to analyze and summarize the results of the included studies. A systematic review differs from a traditional literature review in that a literature review only describes and appraises previous work, and but does not specify the methods by which the reviewed studies were identified, selected, or evaluated (Rychetnik, Hawe, Waters, Barratt, & Frommer, 2004).

tailored messages

Tailored messages include abstracts, full systematic reviews, and evidence summaries. Evidence from health communication studies indicate that computer-tailored communications are associated with increased uptake compared to standardized messages (Suggs, 2006) & and that electronic targeted messaging to subgroups with common interests is effective in promoting evidence-informed decision making (Russell, Greenhalgh, Boynton, & Rigby, 2004). While the provision of short summaries with easy access to full text has been shown to improve uptake of systematic reviews (Lavis, Davies, & Gruen, 2006) and reports (Canadian Health Services Research Foundation, 2001), many questions remain unanswered in terms of what content should be tailored, and what the most effective communication channels are (Suggs, 2006). A paper on a study of tailored messages (A knowledge transfer strategy for public health decision makers), is available on our site.

weak

All reviews in Health Evidence are assessed for methodological quality. Reviews receive a quality assessment score out of 10. Reviews with a score of four or less are rated as weak.

Bandolier. (n.d.). Qualitative and quantitative research. Retrieved from http://www.bandolier.org.uk/booth/glossary/qualres.html.
Barlow Pugh, M et. al.,(Ed). (2006). Stedman’s medical dictionary. Baltimore, MD. Lippincott Williams & Wilkins. ISBN 1071-3964. STAT!Ref Online Electronic Medical Library. Retrieved from http://online.statref.com/document.aspx?fxid=8&docid=38668.
Barratt, A., Wyer, P.C., Hatala, R., McGinn, T., Dans, A.L., Keitz, S., ...Guyatt, G. (2004). Tips for learners of evidence-based medicine: 1. Relative risk reduction, absolute risk reduction and number needed to treat. CMAJ, 171(4), 353-358.
BMJ Publishing Group Ltd & RCN Publishing Company Ltd. (2009). Glossary. Evidence-Based Nursing, 128(4), doi:10.1136/ebn.12.4.128-b
BMJ Publishing Group. (2011). Clinical evidence: Glossary. Retrieved from http://clinicalevidence.bmj.com/ceweb/resources/glossary.jsp.
Buckingham, J., Fisher, B., & Saunders, D. (2008). Evidence based medicine toolkit: Glossaries – clinical epidemiology. Retrieved from http://www.ebm.med.ualberta.ca/Glossary.html.
Calgary Health Region. (2006, August). Knowledge use in the Calgary Health Region: A scan of initiatives that support use of evidence in practice, Glossary of Terms. Retrieved March 4, 2008, from http://www.calgaryhealthregion.ca/programs/research/pdf/knowledge_utilization_scan_full_report_%20aug_28_2006.pdf.
Canadian Health Services Research Foundation. (2001). Communication notes: Reader-friendly writing – 1. 3:25.
Canadian Health Services Research Foundation. (n.d.). Glossary of knowledge exchange terms as used by the Foundation. Retrieved from http://www.chsrf.ca/keys/glossary_e.php.
Centre for Evidence-Based Medicine, University of Oxford. (2005). Critical appraisal for therapy articles. Retrieved from http://www.cebm.net/wp-content/uploads/2014/04/RCT_Appraisal_sheets_2005_English.doc.
The Cochrane Collaboration. (2010). Glossary. Retrieved from http://community.cochrane.org/glossary.
Davies, H.T., & Crombie, I.K. (2009). What is…? Series: What are confidence intervals and p-values? Retrieved from http://www.bandolier.org.uk/painres/download/What%20is%202009/What_are_conf_int.pdf.
The Determinants of Research Utilization: Pain Management in Infants & Children. (2000, April). Glossary. Retrieved from http://www.nursing.ualberta.ca/kusp/rustudy2/glossary.htm.
DiCenso, A. (2001). Clinically useful measures of the effects of treatment. Evidence Based Nursing, 4, 36-39.
DiCenso, A., Guyatt, G., & Ciliska, D. (2005). Evidence-based nursing: A guide to clinical practice. St. Louis, MO: Mosby.
European Observatory on Health Systems and Policies. Glossary. Retrieved from http://www.euro.who.int/observatory/Glossary/TopPage?phrase=effectiveness.
The Evidence for Policy and Practice Information and Co-ordinating Centre (EPPI-Centre). (2006). Glossary. Retrieved from http://eppi.ioe.ac.uk/cms/Default.aspx?tabid=334.
Gay, J. (2010). Clinical epidemiology & evidence-based medicine glossary: Clinical study design and methods terminology. Retrieved from http://www.vetmed.wsu.edu.libaccess.lib.mcmaster.ca/courses-jmgay/GlossClinStudy.htm.
Green, S., & Higgins, J. (Eds). Cochrane handbook for systematic reviews of interventions. Version 4.2.5 [updated May 2005]. Retrieved from htthttp://training.cochrane.org/handbook.
Harvard School of Public Health. (2004). The public health disparities geocoding project monograph glossary. Retrieved from http://www.hsph.harvard.edu/thegeocodingproject/webpage/monograph/glossary.htm.
Higgins, J.P.T., & Green, S. (Eds). (September 2009). Cochrane Handbook for Systematic Reviews of Interventions. Version 5.0.2. [updated September 2009]. Retrieved from http://training.cochrane.org/handbook.
Higgins, J.P.T., Thompson, S.G., Deeks, J.J., & Altman, D.G. (2003). Measuring inconsistency in meta-analysis. British Medical Journal, 327, 557-560.
Lavis, J., Davies, H., & Gruen, R.L. (2006). Working within and beyond the Cochrane Collaboration to make systematic reviews more useful to healthcare managers and policy makers. Healthcare Policy, 1(2), 21-33.
Law, K., & Howick, J. (2011). Centre for Evidence Based Medicine: Glossary. Retrieved from http://www.cebm.net/glossary/.
Leung, W.C. (2000). Balancing statistical and clinical significance in evaluating treatment effects. Postgraduate Medical Journal, 77(905), 201–204.
Mobilizing for Action through Planning and Partnership (MAPP). MAPP glossary. Retrieved from http://www.naccho.org/uploads/downloadable-resources/Programs/Public-Health-Infrastructure/MAPP-Glossary.pdf.
Moher, D., Jadad, A.R., & Klassen, T.P. (1998). Guides for reading and interpreting systematic reviews: III. How did the authors synthesize the data and make their conclusions? Archives of Pediatric and Adolescent Medicine, 152, 915-920.
National Collaborating Centre for Methods and Tools. (n.d.). Online learning module: Introduction to evidence-informed decision making. Retrieved from http://www.nccmt.ca/en/modules/eidm/pdfs/EIDMmodule_define_en.pdf.
National Health Service. (2009). Glossary. Retrieved from http://www.nhs.uk/news/Pages/Newsglossary.aspx.
National Institute for Health and Clinical Excellence. (2010). Glossary. Retrieved from http://www.nice.org.uk/website/glossary/.
O’Donnell, M.P. (1989). Definition of health promotion: Part III, expanding the definition. American Journal of Health Promotion, 3(5).
Organization for Economic Co-Operation and Development, Development Assistance Committee (DAC) Working Party on Aid Evaluation. (2002). Glossary of key terms in evaluation and results based management. Retrieved from http://www.oecd.org/dataoecd/29/21/2754804.pdf.
Public Health Agency of Canada. (2002). What is the population health approach? Retrieved from http://www.phac-aspc.gc.ca/ph-sp/approach-approche/index-eng.php
Public Health Agency of Canada. (2006, October). Glossary of terms relevant to the core competencies for public health, E-H. Retrieved from http://www.phac-aspc.gc.ca/ccph-cesp/glos-e-h-eng.html.
Public Health Agency of Canada. (2006, October). Glossary of terms relevant to the core competencies for public health, I-P. Retrieved from http://www.phac-aspc.gc.ca/ccph-cesp/glos-i-p-eng.html.
Russell, J., Greenhalgh, T., Boynton, P., & Rigby, M. (2004). Soft networks for bridging the gap between research and practice: Illuminative evaluation of CHAIN. BMJ, 328, 1174-7.
Rychetnik, L., Hawe, P., Waters, E., Barratt, A., & Frommer, M. (2004). A glossary for evidence based public health. Journal of Epidemiology and Community Health, 58, 538-545.
Sackett, D.L., Richardson, W.S., Rosenberg, W., & Haynes, R.B. (1997). Evidence-based medicine: How to practice and teach EBM. New York: Churchill Livingstone.
Snowden, D. (2009, September 23). Defining KM [Web log comment]. Retrieved from http://www.cognitive-edge.com/blogs/dave/2009/09/defining_km.php#more.
Stanhope, M., & Lancaster, J. (1996). Community health nursing: Promoting health of aggregates, families, and individuals. Toronto: Mosby.
Suggs, L.S. (2006). A 10-Year retrospective of research in new technologies for health communication. Journal of Health Communication, 11, 61-74.
SUPPORT. (n.d.). Glossary. Retrieved from http://www.support-collaboration.org/summaries/c.htm.
Waters, E., Doyle, J., Jackson, N., Howes, F., Brunton, G., & Oakley, A. (2006) Evaluating the effectiveness of public health interventions: The role and activities of the Cochrane Collaboration. Journal of Epidemiology and Community Health, 60, 285-289.
World Health Organization. (1998). Health promotion glossary. Retrieved from http://www.who.int/healthpromotion/about/HPR%20Glossary%201998.pdf.
World Health Organization. (2006). Public health. Retrieved from http://www.who.int/trade/glossary/story076/en/print.html.
World Health Organization. (2011). Health topics: Public health surveillance. Retrieved from http://www.who.int/topics/public_health_surveillance/en/.