RealRisk: FAQs
Can RealRisk be used with any research paper?
RealRisk can be used with any study which investigates the link between a risk factor and an outcome of interest and reports a
relative risk
In a research study, a risk is the probability of an outcome occurring in one group (e.g. the number participants having a heart attack over the total number of participants).
A relative risk is the risk in the experimental group divided by the risk in a control or baseline group (also sometimes called the ‘Risk Ratio’).
A relative risk greater than 1 means the outcome was more common in the experimental group than the control/baseline group, and a RR less than 1 means it was less common.
Where do I find it?
Relative Risks
will normally be reported in the abstract of a paper and in the results section. They generally look like this:
“RR 3.6”
or sometimes
“aRR 3.6”.
Often, several RRs will be reported for different comparisons. For RealRisk, pick the RR most worth reporting.
(RR),
hazard ratio
A hazard is the rate at which some outcome of interest occurs over a given period of time (e.g. heart attacks or cancer diagnoses per year).
A hazard
ratio
is the hazard in an experimental group (exposed to the risk factor) divided by the hazard in a control or baseline group.
A hazard ratio bigger than 1 means the outcome of interest has occurred at a higher rate in the experimental group than the control or baseline group, and a hazard ratio smaller than 1 means it’s occurred at a lower rate.
Where do I find it?
Hazard ratios
will normally be reported in the abstract of a paper and/or in the results section. They generally look like this:
“HR 3.6.”
. Often, several HRs will be reported for different comparisons. For RealRisk, pick the HR most worth reporting.
(HR) an
odds ratio
In a research study, the odds of some outcome is the number of times it happened over the number of times it didn’t happen (e.g. the number of people who had heart attacks over the number who didn’t)
An odds
ratio
is the ratio of two odds: the odds of the outcome of interest in the experimental group divided by the odds in a control or baseline group.
An odds ratio greater than 1 means the outcome of interest (heart attacks or cancer diagnoses) was more common in the experimental group (usually the group exposed to the risk factor), and an OR below 1 means it was less common.
Odds ratios will normally be reported in the abstract of a paper and/or in the results section. They generally look like this:
“OR 3.6”
or sometimes
“aOR 3.6”.
Often, several ORs will be reported for different comparisons. For RealRisk, pick the OR most worth reporting.
(OR) or a
percentage change
This refers to a percentage increase or decrease as in, “Women taking HRT were 80% more likely to develop breast cancer” or “People who exercise are 70% less likely to be depressed”.
For a percentage increase just type in the number e.g. 80%
To indicate a percentage decrease add a minus sign, e.g -70%
. If none of these statistical figures are reported, the paper is not suitable for RealRisk.
Here are some examples of papers that would be suitable for RealRisk:
- Type and timing of menopausal hormone therapy and breast cancer risk: individual participant meta-analysis of the worldwide epidemiological evidence.
- Consumption of sugar‐sweetened and artificially sweetened soft drinks and risk of cancers not related to obesity.
- Association between Excessive Use of Mobile Phone and Insomnia and Depression among Japanese Adolescents.
What if I don't know all the inputs?
Only the inputs ‘paper title’ and ‘DOI’ are optional. The essential inputs for the calculation are baseline risk and change in relative risk (RR, HR, OR). The rest of the inputs requested are labels explaining what those numbers mean, but they are crucial for producing a meaningful result, that is why they do not appear as optional. All of these inputs should be in the paper, but sometimes, particularly the baseline risks are not reported. Guidance is provided in the next question.
What if the baseline risk is not in the paper?
Check the paper carefully: introduction, methods and results are the most common sections to include the baseline risk used in the study. If it is not in the paper, here are some alternatives:
- Calculate the percentage yourself.
Sometimes the raw numbers are reported and you could calculate the percentage yourself (eg. ‘In the group receiving placebo 6 out of 134 people developed antibiotic resistance’, then, the baseline risk would be 4%).
- Request it from the corresponding author
Contact the corresponding author and request it. Remember, if their study reports an RR, OR or HR they are explicitly comparing outcomes between two groups. You are just asking for the risk (or probability) of the outcome in one of those groups - the control or baseline. (You could even ask them to fill out RealRisk themselves!)
Researchers may point out that these numbers sometimes come out of statistical models, and don’t emerge straightforwardly from the raw numbers. That is true, but don’t be fobbed off - they can still calculate the baseline risk!
- Find a relevant figure from a trustworthy authority.
If the corresponding author is unhelpful, you may have to look elsewhere for a relevant statistic. Charities like
,
,
etc. may be able to help, and their websites often include the relevant facts and figures. Other useful organisations include the
, the
for US stats, and the
or
for the UK.
If you are a journalist looking report a piece of research, note that you don’t have to use exactly the same baseline risk as the one used in the paper
For example, if a piece of research reports a 70% increased risk (an RR of 1.7) of depression amongst under 18s year olds who vape, you could quote a trusted authority on the percentage of under 18s who suffer from depression – and then apply the 70% increase to that. It may not be the exact baseline figure used in the study, but if it’s a trustworthy number with direct relevance to your audience, it’s a correct and effective way to report the research.
What if the paper includes multiples risk results? How do I know what numbers to use?
It is common that papers report multiple results because, for example, they could have tested the effect of multiple doses, or different foods or different treatments. RealRisk can only analyse results one at a time. So, select one result to input, probably the one with the biggest effect in the study, and you could do the same with as many results as you want. Any time you input a different result you may have to change other inputs.
Who designed the website?
At the Winton Centre we believe that everyone has a right to balanced evidence on issues important to them; evidence presented in a transparent way, to inform but not persuade.
We aim to seek out the evidence about issues important to people’s lives, and ensure that it is presented to them in such a way as to make the potential for both risk and benefit as clear as possible. We want to encourage global adoption of the best-known methods of communicating quantitative evidence clearly and without bias.
If you want to get in touch about RealRisk - to share feedback or if you need some help - email
wintoncentre@maths.cam.ac.uk Where can I find more information on how to report risk and statistics?
A podcast presented by David Spiegelhater that explores how risk should be communicated in areas like health, environment and politics.
by the Science Media Centre
by Sense About Science, 2010
