meta-analyses / meta-analysis-shiny Goto Github PK

Reorder dose before person years

• Replace both by fatal-and-non-fatal
• Replace Sub-population by Sex-stratified
• Introduce two levels to select outcomes
• Full form of acronyms
• Relabel data table columns
• Update text of PIF tables
• Update figure labels

Both DR and individual studies download buttons need to be reactive. Also remove global variables.

we need to calculate population impacts fractions for each of the diseases- see attached
PIF.xlsx

in the title of each graph pleaes can we see ns for studies & people (or person years)

Represent all three positions

Along with depressive symptoms and major depression

outcome type , population and last knot in a single column

CVD Cardiovascular Disease
CHD Coronary Heart Disease

in the side bar and in graph titles

in graph please can we have Marginal MET hours per week

If the RR does not go above 1 please can we see from 0 to 1 only on Y axis.

could it include
a) links to github code including Alessio's code
b) explanation of the quantiles (to be written by Alessio or Nicola)
c) short description of the PA data- to be written by Soren
d) a few warnings !

so we can have (as an option with radio button) to see all the studies overlayed on top of the main curve. For these studies the size of each point should be weighted by the inverse of the confidence interval (apart from for reference category ie. not 1/0!)

Update download data when there is no data

Please can we see y axis with more points & have ability to zoom into one plot so we can see more precisely the values.
The x axis is better but could also do with more points

Also make sure that the knots labels are at the correct position when y-axis log transformation happens.

it is hard to select the CI on the top because the tool tip from the central estimate covers the point you want to select

Make both PIFs and RRs tables collapsable for more space

For log y-axis, sometimes the value 1 is not selected. Make sure to use 1 in all cases.

From James:

can we have flexible choice of PI with it showing what % of the population have exposure above that level.

with it showing what % of the population have exposure above that level.

Add summary and plot

In order to deal with the error:

length of 'center' must equal the number of columns of 'x'

remove problematic studies

Remove study # 70 from strokes and study #38 from CHD