Add diff_est_ci statistic to s_proportion_diff()

[vikram-rawat]

Feature description
s_proportion_diff() currently returns diff and diff_ci as
separate elements. This means the risk difference and its CI always
render as two table rows.

A combined 3-element vector diff_est_ci = c(diff, lower_ci, upper_ci)
would allow rendering them on a single row (e.g., -7.3 (-26.7, 12.2)).

This pattern already exists in tern — s_surv_timepoint() returns
rate_diff_ci_3d for the same purpose.

Fixes - #1523

[munoztd0]

minor change but overall lgtm

[munoztd0]

all good now

[wwojciech]

@Melkiades , @shajoezhu - can you pl. have a look at this PR and possible merge it if you have no any objections? Thank you!

[shajoezhu]

@munoztd0 , can you share the scda.test test results? PR link here?

@wwojciech can you resolve the conflict?

[wwojciech]

@vikram-rawat - can you resolve conflicts please? I unfortunately have no permissions to do that. Thanks!

[shajoezhu]

@wwojciech @gmbecker , could you give me a list of members that i will add you to the team for access. thanks

[wwojciech]

@wwojciech @gmbecker , could you give me a list of members that i will add you to the team for access. thanks

Thank you @shajoezhu - I will check this with the team and will get back to you. In the meanwhile, for sure you can grant me an access. Thanks!

[munoztd0]

@munoztd0 , can you share the scda.test test results? PR link here?

@wwojciech can you resolve the conflict?

OFC, -> here it is -> insightsengineering/scda.test#251

[munoztd0]

@wwojciech @gmbecker , could you give me a list of members that i will add you to the team for access. thanks

@gmbecker @wwojciech @vikram-rawat @iaugusty @eanokian @rmao6 @y-runthla and @danielinteractive @gm-gi-ext

[wwojciech]

Looks ok.

[Melkiades]

Approving. Two small things I pushed straight to the branch rather than send you round again:

• tern_default_labels[["diff_est_ci"]] hardcoded "... and 95% CI". get_labels_from_stats() is exported and returns that value, so it was wrong for any other conf_level. Now conf-level agnostic; the label attribute from s_proportion_diff() still carries the actual level, so rendered output is unchanged.
• NEWS entry now carries (#1523), matching its neighbours.

Checked after the change: default table output identical, diff_est_ci renders correctly at both 95% and 90%, test-prop_diff.R 89 passed and 0 failed.

One thing I deliberately left alone: .indent_mods lists diff_est_ci but .formats does not. "xx.x (xx.x, xx.x)" is not a built-in formatters spec, the built-in 3d ones use " - " rather than ", ", so it needs format_xx(), and putting a function into that vector would turn .formats into a list. The default lookup in tern_default_formats already resolves it, so the asymmetry is harmless.

Nice addition, the single-row rendering is genuinely useful.