Forest plot

A meta-analysis at a glance: each study's estimate as a square sized by weight with its confidence interval, a line of no effect, the pooled estimate as a diamond, and the numbers in columns alongside. Ratios read on a log axis.

Stroke within five years

Odds ratios from seven placebo-controlled trials, pooled by inverse-variance weight.

Alvarez 20110.72 [0.55, 0.94]18.2%

Brennan 20130.85 [0.62, 1.17]12.5%

Chowdhury 20140.64 [0.45, 0.91]10.1%

Duarte 20160.91 [0.74, 1.12]24.3%

Eriksen 20180.58 [0.36, 0.93]6.4%

Fujimoto 20190.79 [0.63, 0.99]20.6%

Gallagher 20211.08 [0.71, 1.64]7.9%

Pooled (fixed effect)0.80 [0.72, 0.89]100%

Odds of a stroke within five years, treatment against placebo
Study	Odds ratio [95% CI]	Weight
Alvarez 2011	0.72 [0.55, 0.94]	18.2%
Brennan 2013	0.85 [0.62, 1.17]	12.5%
Chowdhury 2014	0.64 [0.45, 0.91]	10.1%
Duarte 2016	0.91 [0.74, 1.12]	24.3%
Eriksen 2018	0.58 [0.36, 0.93]	6.4%
Fujimoto 2019	0.79 [0.63, 0.99]	20.6%
Gallagher 2021	1.08 [0.71, 1.64]	7.9%
Pooled (fixed effect)	0.80 [0.72, 0.89]	100%

Usage

use hydroxy_leptos::chart_frame::ChartFrame;
use hydroxy_leptos::forest_plot_chart::*;
use leptos::prelude::*;

/// Seven trials of a blood pressure drug against placebo: odds ratio of a stroke within
/// five years, its 95% interval, and the trial's inverse-variance weight.
fn trials() -> (Vec<Study>, Study) {
    let studies = vec![
        Study::new("Alvarez 2011", 0.72, (0.55, 0.94), 18.2),
        Study::new("Brennan 2013", 0.85, (0.62, 1.17), 12.5),
        Study::new("Chowdhury 2014", 0.64, (0.45, 0.91), 10.1),
        Study::new("Duarte 2016", 0.91, (0.74, 1.12), 24.3),
        Study::new("Eriksen 2018", 0.58, (0.36, 0.93), 6.4),
        Study::new("Fujimoto 2019", 0.79, (0.63, 0.99), 20.6),
        Study::new("Gallagher 2021", 1.08, (0.71, 1.64), 7.9),
    ];
    (studies, Study::new("Pooled (fixed effect)", 0.80, (0.72, 0.89), 100.0))
}

pub fn demo() -> AnyView {
    let (studies, pooled) = trials();
    view! {
        <ChartFrame title="Stroke within five years" description="Odds ratios from seven placebo-controlled trials, pooled by inverse-variance weight.">
            <ForestPlotChart
                studies
                pooled=Some(pooled)
                label="Odds of a stroke within five years, treatment against placebo"
                measure="Odds ratio"
                log=true
                favours=("Favours treatment".to_string(), "Favours placebo".to_string())
            />
        </ChartFrame>
    }
    .into_any()
}

use dioxus::prelude::*;
use hydroxy_dioxus::chart_frame::ChartFrame;
use hydroxy_dioxus::forest_plot_chart::{ForestPlotChart, Study};

/// Seven trials of a blood pressure drug against placebo: odds ratio of a stroke within
/// five years, its 95% interval, and the trial's inverse-variance weight.
fn trials() -> (Vec<Study>, Study) {
    let studies = vec![
        Study::new("Alvarez 2011", 0.72, (0.55, 0.94), 18.2),
        Study::new("Brennan 2013", 0.85, (0.62, 1.17), 12.5),
        Study::new("Chowdhury 2014", 0.64, (0.45, 0.91), 10.1),
        Study::new("Duarte 2016", 0.91, (0.74, 1.12), 24.3),
        Study::new("Eriksen 2018", 0.58, (0.36, 0.93), 6.4),
        Study::new("Fujimoto 2019", 0.79, (0.63, 0.99), 20.6),
        Study::new("Gallagher 2021", 1.08, (0.71, 1.64), 7.9),
    ];
    (studies, Study::new("Pooled (fixed effect)", 0.80, (0.72, 0.89), 100.0))
}

#[component]
pub fn ForestPlotChartDemo() -> Element {
    let (studies, pooled) = trials();
    rsx! {
        ChartFrame { title: "Stroke within five years", description: "Odds ratios from seven placebo-controlled trials, pooled by inverse-variance weight.",
            ForestPlotChart {
                studies,
                pooled: Some(pooled),
                label: "Odds of a stroke within five years, treatment against placebo",
                measure: "Odds ratio",
                log: true,
                favours: ("Favours treatment".to_string(), "Favours placebo".to_string()),
            }
        }
    }
}

Anatomy

Component	data-part	Renders
`ForestPlotChart`	root	figure; `data-active` while a row is picked out
	plot	Focusable grid of rows; arrow keys step through the studies, then the pooled row
	head	Column headings
	row	One study, or the pooled result with `data-pooled`; `data-active` when picked out
	study	The study's name
	interval	The plot cell holding the marks
	square	The estimate; its area tracks the study's weight
	whisker	The confidence interval; extends out from the estimate
	diamond	The pooled estimate and interval; lands last
	estimate	Estimate and interval as text
	weight	The study's share of the total weight
	null	The line of no effect: 1 on a log axis, else 0
	axis	Ticks under the plot column, with the favours labels either side of no effect
	tooltip	Estimate, interval and weight of the active row
	sr-table	Every row with its estimate, interval and weight

Keyboard

Key	Does
`Arrow keys`	Step through the rows
`Home / End`	First or last row
`Escape`	Clear the active row

Props

Prop	Type	Notes
`studies`	Vec<Study>	Required. Study::new(label, estimate, (low, high), weight)
`pooled`	Option<Study>	The pooled result, drawn as a diamond; its weight is ignored
`label`	String	Required. What the chart shows; the accessible name
`measure`	String	Required. The effect measure, such as "Odds ratio"
`log`	bool	Log axis with no effect at 1, for ratios. Default false (no effect at 0)
`favours`	(String, String)	What each side of no effect means, shown under the axis
`level`	u8	Confidence level in the column heading. Default 95
`format`	Fn(f64) -> String	Formats estimates, limits and ticks. Default: two decimal places

Events

Prop	Type	Notes
`on_hover`	Callback<Option<Datum>>	The mark under the pointer or keyboard focus changes, or is left. Fires only on a change
`on_select`	Callback<Datum>	Click, tap or Enter on a mark