Questions and lessons can't be printed or saved as a PDF.
Which part of the forest plot carries this information?
More, if you want it
Which mark on the chart is the patient asking about?Tap a term: its definition, what you can do with it, and how often the exam asks it.
Definition: the block that shows one study's result.
You can: Tell a point estimate from block size by distance from the line of no effect.
Definition: the sideways line that shows the range where the true value probably lies.
You can: Tell a narrow confidence interval from a far point estimate by the width of the line.
Definition: an arrow tip that shows the interval runs off the scale.
You can: Name an arrowhead when a study's line runs off the scale.
Definition: a result whose interval does not cross the line of no effect.
You can: Tell statistical significance from a favourable point estimate by whether the interval crosses the line.
Definition: the upright line at 1 for ratios and at 0 for differences.
You can: Name the line of no effect when choosing 1 or 0 for the measure.
Definition: the combined result of all studies, with its interval as its width.
You can: Tell the pooled diamond from a single study by its shape at the bottom of the plot.
Each card is one term. The grey line is what the patient says, which is what the question stem gives.
Summary
The line of no effect sits at 1 for ratio measures and at 0 for difference measures.
The line of no effect sits at 1 for ratio measures and at 0 for difference measures.
An odds ratio is one odds divided by another. The odds ratio compares the odds of an event in the treated group with the odds in the control group. Odds ratio and relative risk are ratio measures. Mean difference is a difference measure. A ratio uses 1, and a mean difference uses 0. Priya says, "The axis shows odds ratios, so I check the line at 1."
Priya is a core trainee presenting an antidepressant at journal club.
Candidates read every plot against a line at 0. Check the measure on the axis before reading any interval.
The line of no effect sits at 1 for ratio measures and at 0 for difference measures.
The largest effect is the study whose point estimate lies furthest from the line of no effect.
The largest effect is the study whose lies furthest from the line of no effect.
The plot lets the reader compare the size of effect between studies. The study with the largest effect stays the same, whatever its block size or interval width. Each study adds its own odds ratio or effect size to the plot. Block size shows the weight of a study. Tomasz says, "The small block far from the line has the bigger effect."
Tomasz, a registrar, studies a plot with a large block near the line and a small block far away.
The distance from the line of no effect shows the effect size.
Mnemonic
Far from the fence is big, short stick is sure. Cue: shows the largest effect
The fence is the line of no effect, and a study standing far from it has the largest effect. The stick the hiker holds sideways is the confidence interval, and the shortest stick marks the most precise study.
The largest effect is the study whose point estimate lies furthest from the line of no effect.
The most precise study has the narrowest confidence interval, usually the largest study.
The most precise study has the narrowest , usually the largest study.
A narrower interval means a more precise study. A wider interval means less precision. An arrowhead on the line means the interval runs off the plotted scale. The estimate is then very imprecise. Each confidence interval is a horizontal line through its block. A far block can carry a long line. Amara says, "This trial is too small to trust alone."
Amara, a foundation doctor, sees one study line end in an arrow tip.
The shortest line marks the most precise study, wherever its block sits.
Analogy
Each study is a hiker on a path beside a fence. The fence is the line of no effect. The hiker's distance from the fence is the effect. The hiker holds a stick sideways, and the stick is the interval. A stick that reaches over the fence means not significant.
Where the analogy breaks: a hiker's build cannot show the block size, which is the study's weight.
The most precise study has the narrowest confidence interval, usually the largest study.
A study is statistically significant only if its confidence interval does not cross the line of no effect.
A study is statistically significant only if its confidence interval does not cross the line of no effect.
Count significant studies by their intervals. A point estimate on the other side of the line favours the comparator. The interval is the range where the true value probably lies. The further the interval sits from the line, the stronger the significance. The first interval above crosses 1. Kwame says, "Only the second trial is significant."
Kwame, a core trainee, reads one interval that crosses 1 and one that lies wholly below 1.
Candidates count every block on the favourable side as significant. An interval that touches or crosses the line is not significant.
A study is statistically significant only if its confidence interval does not cross the line of no effect.
The pooled result is a diamond whose width is its confidence interval.
The pooled result is a diamond whose width is its confidence interval.
Sofia, a specialty trainee, reads a diamond that lies clear of 0 and close to it.
Candidates judge the overall impression of the diamond by its width alone. A diamond that crosses the line is not significant.
Meta-analysis combines the studies' odds ratios or effect sizes into one pooled estimate. The diamond's distance from the line gives the size, such as small to moderate. Sofia says, "The result is significant, and the effect is small."
The pooled result is a diamond whose width is its confidence interval.
Block size shows weight only.
| Part | Reader looks at | Often mistaken for | Deciding feature |
|---|---|---|---|
| Point estimate | Block position | Block size | Distance from the line |
| Confidence interval | Line width | Effect size | Narrowest is most precise |
| Statistical significance | Interval and line | Favourable side | Interval clear of the line |
| Pooled diamond | Bottom shape | Single study | Crossing and distance |
| Line of no effect | Axis measure | Always 0 | 1 for ratios, 0 for differences |
| Arrowhead | Line ends | Large effect | Interval runs off the scale |
Block size shows weight only. Effect is distance, is width, and significance is clearance from the line.
Block size shows weight only.
mrcpsych.co.uk · An independent study tool · not affiliated with the Royal College of Psychiatrists